§01 · Changelog

Every release, in plain language.

We ship improvements every few weeks, driven by real scan accuracy audits and user feedback. No marketing fluff, just what changed.

v4.58August 18, 2026Latest

Every tag in the AI Radar movement table now explains itself. Climbing, Slipping, New, Returned, Dropped, Steady and Not cited each carry a definition on hover or tap, so a row can be read without leaving the table.
A query that came back to the scan after a gap now reads honestly. Returned means the query is being checked again, not that a citation was won, so a returned query with no citations sits in neutral grey and says not cited instead of reading as a gain.
Those definitions follow the period you are looking at, so the month and quarter views describe a month or a quarter rather than a week.

v4.57August 18, 2026

Opening a page workspace from your dashboard now works for a site's homepage. Homepage links resolved to a not-found page even when the page had been scanned; every page you have scanned now opens its workspace.
The page trend chart reads correctly when every scan happened on the same day. One batch of scans is one measurement, so it now sits in a single column with the week ahead left open, instead of being spread across the chart as though the pages were scanned days apart.
When each page has only been scanned once, the chart says so, and the space to the right is the wait until your next scheduled scan, marked on the axis with its date. Clients without a schedule keep a plain one-week runway and no marker.
Every page workspace now says it is a summary of your latest scan and carries an Open full report button at the top, so the visitor breakdown, exit reasons and pillar scores behind the summary are always one click away.

v4.56August 17, 2026

The free AI Radar check now says it takes about 3 minutes, because that is what it measures. Every real check completed in under two and a quarter minutes, and the progress bar is paced to match, so it reads as nearly there rather than time to go and do something else.
Client reports are easier to read. Every muted line in the report and in the site audit now meets the WCAG AA contrast standard on both the web version and the PDF, so the smaller supporting text is legible rather than faint.
AI Radar's teal is deeper, which lifts the active tab, the period chips and the query filters clear of the AA threshold against their own tinted backgrounds.
Status colours across every report, email and share card now use the calibrated values, so an amber warning or a red loss reads the same wherever it appears.
Keyboard navigation is tidier. The carousel dots and the cropped report preview on the AI Radar pages no longer take keyboard focus, and the target market picker in Client settings now announces its name to a screen reader.

v4.55August 14, 2026

The AI Radar performance chart now plots every AI engine as its own line, alongside the all-engine average. One engine carrying you while another never cites you is visible at a glance instead of averaging away.
Click any engine name above the chart to hide or show its line, so you can isolate two engines and compare them directly. Hovering a week now lists that week's number for each engine.
An engine line breaks where that engine was not run that week rather than dropping to zero, so a skipped scan never reads as a lost citation.

v4.54August 11, 2026

New blog post: what each AI engine actually wants before it cites you. A per-engine breakdown of ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews, covering what each one rewards, what silently blocks you, and which popular AEO advice is now out of date.

v4.53August 11, 2026

AI Radar now reports how AI answers describe your brand. Each answer that names you is classified for the qualities it attaches (trusted, premium, expensive, complex to set up, and more), shown as chips on every engine result and rolled up on the Appearance view.
The Appearance view gains a recommendation rate: on vendor-comparing and ready-to-buy questions, the share of answers that actually recommended you rather than mentioning you in passing.
Every question on the Coverage report carries a stability read (Stable, Shifting, or Volatile) derived from repeat runs, cross-engine agreement, and recent weeks, so you can tell a real gap from engine noise.
When answers keep recommending you with caveats, the fix on your Action plan now names the recurring objection, such as price or setup complexity, instead of asking you to find it yourself.
The AI Radar pages now explain the six measurement signals behind every weekly scan: intent coverage, recommendation rate, attribute framing, source dependency, stability, and gap mapping.

v4.52August 11, 2026

The review prompt now asks once per product and remembers your answer. Decline it and it stays away; leave a review and we never ask again on either product.
AI Radar has its own review prompt, shown on a week where AI actually cited you or right after you publish a client report, rather than at a moment where there is nothing to say.
Opening a single page from inside a site scan now carries the full review flow, including the Trustpilot link.

v4.51August 10, 2026

Leaving a review is now much quicker. The review prompt on your report routes to Trustpilot, where a review takes about a minute with a Google or email sign-in, instead of platforms that required business-email or LinkedIn verification.
Your star rating carries over automatically, and the sentence you already wrote in the report is shown with a copy button, so there is nothing to compose twice.
Review links in the scan-complete and follow-up emails now always point somewhere real, with the same one-minute Trustpilot flow.

v4.50August 10, 2026

Your first AI Radar scan now runs the moment you subscribe or redeem a code, instead of waiting for the next scheduled one. We email you the results as soon as they land, usually within a few minutes.
That first scan runs at your plan's full depth, so it counts as a real week on your trend rather than a lighter preview sitting next to your proper scans.
Weekly scans are now set up for you automatically, on the weekday you signed up. You can still change the day and time whenever you like in Settings.

v4.49August 10, 2026

AI Radar access codes have their own page at /radar/redeem. Enter a code and your plan is set for a fixed free period, with no card taken and no payment form to work past.
If you subscribe while a free period is running, you keep the rest of it. Your first payment is scheduled for the day the free period ends, and the pricing page shows you that date instead of the price.
A bar across your Radar workspace shows how long is left on a free period, and we email you a week before it ends, the day before, and on the day.
When a free period ends, nothing is deleted. Your tracked questions, prompt bank and every week of history stay exactly as they were.
Redeeming a code sets your scan running straight away, and a welcome email confirms what is happening, so you are never left wondering whether it worked.

v4.48August 9, 2026

The AI Radar free check now opens in the real Radar workspace. Your result renders through the same verdict card, movement table, competitor read, and action plan a paying account sees, with the same sidebar, so what you review is the product itself rather than a one-off report page.
Every part of the workspace is explorable from a free check. This month, Analytics, and Client reports open with an honest note on what unlocks with weekly tracking, plus the real product views on clearly labeled sample data, so you can see exactly what you would get before paying.
If you subscribe after a free check, the check's results carry into your new account as week one. Your questions become tracked prompts, the trend chart starts from your check instead of from zero, and your first weekly scan becomes point two on the line.
The results email now reads like the weekly scan email the paid product sends: the same cards for what we found, which engines named you, and who came up instead, plus a preview of what the weekly version adds.
Signing in with a live free check now lands you back on your own results instead of the homepage.
There is a sample of the new free check experience at /radar/check/demo.

v4.47August 9, 2026

Regenerating a client's prompts in AI Radar no longer removes the prompts you already had. They move below the scan line and stay on the Prompts page, so anything you had curated is one drag away from being scanned again instead of something you have to type back in.
If the regenerated bank happens to ask a question you were already tracking, that question keeps its existing row and its full week-by-week history rather than starting over as a duplicate.
The same applies to regenerating brand-defense queries.
Appending new category queries now skips any the bank already tracks, so a question cannot end up scanned twice a week with its history split across two rows.

v4.46August 9, 2026

The email field in the site footer now shows what you type. It was inheriting colours from our previous dark design, which left the text almost the same shade as the background it sat on.
Email preference labels in Settings are readable again, for the same reason.
Improved text contrast across pricing, the blog index, the dashboard and Settings. Small labels, captions and status chips are now comfortably legible rather than technically present, and warning and error text is darker so it reads clearly on the light background.

v4.45August 9, 2026

In AI Radar, the stored answer under "What ChatGPT actually said" now reads the way the engine wrote it. Headings, numbered steps, bullet lists and comparison tables keep their shape instead of being flattened into one block of text, so a long answer is something you can scan rather than something you have to decipher.
Links inside a stored answer are clickable, and the reference markers that pointed into a numbered source list we do not show you are gone, so the quote reads as a quote.

v4.44August 8, 2026

Page speed is now judged more conservatively. Our scanner loads your page from Europe on shared infrastructure, so part of every reading was our distance and our load rather than your site. Speed findings now account for that, which means a page that is fast for your visitors is no longer reported as slow.
Reports no longer print a raw load-time figure. A number measured from one location is not one we can stand behind for your visitors, so page speed now reads as a plain status and points you to Core Web Vitals in Search Console, which reflects where your visitors actually are.
Simulated visitors only react to load speed when the delay is genuinely your page's own, so speed stops crowding out the fixes that matter more on pages where it was never the real problem.

v4.43August 8, 2026

New AI Radar setup guide at /docs/ai-radar/setup: six steps from checkout to your first weekly read, illustrated with real screenshots of the actual workspace on every step.
The guide walks the parts that matter most: checking your target market (new purchases default to the United States), reviewing and reordering the generated prompts, the weekly scan line, awareness stage, and how to read your first This Week screen.
The guide is linked wherever you'd look for it: the AI Radar menu, the Set-it-up-once section on the AI Radar page, the how-it-works walkthrough, the pricing cards (setup takes about ten minutes), the agencies page, and the docs hub.

v4.42August 8, 2026

Starting a scan from inside a client now pre-fills the URL field with that client's main site, so a rescan is open and go. The field is still fully editable if you want a different page.
Page headings in a site report now say Full page report next to them, so it is obvious you can open any page for its complete visitor breakdown. The ranked page chart is clearly clickable too.
The homepage now reads as / (Homepage) throughout a report instead of a bare slash, so it is easy to pick out in a list of paths.
Bounce now reads as a likelihood of High, Moderate or Low rather than a predicted percentage. The underlying figure was always a count of how many of the 50 simulated visitors left, never a forecast of your real traffic, and the wording now says exactly that.
Click Intent is now called CTA Pull on every surface and reads as a score out of 100, matching how the metric is described everywhere else in the product.

v4.41August 7, 2026

The AI Radar page now shows a real 14-week trend instead of six points, so the shape of a climb is visible: a flat, noisy start, then the line lifting once the off-page work lands.
Hovering any week now reads as citations against the number of checks behind it, such as 6 citations from 40 checks, rather than a bare count of answers.
The whole sample account was rebuilt at the volume a real client is measured at: 8 buyer questions across all 5 engines, 40 answers checked every week. The weekly movers, the monthly rollup, and the sample client report all derive from those same weeks, so the numbers agree wherever you look.

v4.40August 7, 2026

The nav entry into your workspace now reads Launch App, so it is obvious which link opens the product and which ones are the marketing pages.
Signed in and clicking AI Radar then Overview now takes you to the overview page. It used to bounce you to your dashboard, which meant customers could not reach their own product's pricing, comparisons, or feature pages.
The AI Radar overview also loads faster now that it no longer checks who you are before rendering.

v4.39August 7, 2026

The AI Radar page now previews the current workspace exactly as it ships: the This Week view with its movement leaderboard (which prompts climbed, slipped, or dropped, with per-prompt trend sparklines) and the this-week-vs-last coverage bars.
A new This Month view joined the preview: the month-over-month delta, the week-by-week trajectory line with its plain-English read, and how many of your questions were cited at all.
The preview now tours the workspace on its own, stepping through This Week, This Month, Prompts, the per-engine deep-dive, and the white-label client report. It stops the moment you interact, and respects reduced-motion settings.
The preview's navigation now matches the live app exactly, so what you see before buying is what you get after.

v4.38August 7, 2026

A free scan now asks for your email alongside the URL, and the report is saved to that address the moment it finishes. No account to create, no password to pick.
Your report arrives by email with a one-click link that opens it signed in and drops it into your dashboard, so later scans of the same page are compared against it. The link keeps working for 30 days and on any device.
Come back with an address you have already scanned with and we take you to your dashboard rather than turning you away. Everything you have run is still there.
We only email you about the scan you asked for. No newsletter, and nothing else switched on by default.

v4.37August 7, 2026

New post: Why AI Doesn't Cite You. Most AEO advice is a list of things that earn a citation. This one covers the five structural blockers that prevent one, engine by engine, including the crawler that actually gates each of ChatGPT, Perplexity, Claude, Google AI Overviews and Gemini.
It opens with our own failure: a robots.txt that allowed every AI crawler by name, while our CDN returned a 403 at the edge before any of them reached the file. Nothing errored, and citation just stopped.
It also defuses the Cloudflare 15 September change rather than adding to the panic, and covers the Brave index, which gates a large share of Claude's citations and has almost no competition on it.

v4.36August 4, 2026

Crawler-access fixes now come with the exact robots.txt lines to paste, listing only the AI search crawlers your site is actually missing. They also remind you to check your CDN's bot settings, because the edge can refuse a crawler your robots.txt welcomes.
Schema fixes now include a ready-made Organization block built from your own site name and the profiles we detected, rather than a template with blanks in it.
Colour contrast failures now propose a specific colour that passes, checked against the exact background we measured on your page. Where no colour in that hue can pass, we say so instead of suggesting one that still fails.
Every statistic behind a recommendation now carries a date and a source, and stops being quoted once it is too old to trust. The advice stays; the stale number does not.

v4.35August 4, 2026

Fixes that come with ready-to-paste code now carry a copy button wherever you meet them, not just on the report page. The tag block, the canonical link, the FAQ markup: grab it from your dashboard fix queue and paste it straight into your site.
Your fix queue now names the page element each conversion fix is about, so you know where to look before you open the report.
Page scan CSV exports now include the full fix list across all four scores, matching what site scans already gave you. Paste it into your tracker and work down the list.
Markdown exports now carry the AI Citability, Accessibility and Search Rank fixes, with the code blocks intact.
The AI Radar action plan can now be exported as CSV or Markdown, with a checkbox per action and the measured evidence behind each one.
Client reports now show the numbers behind each recommended action, so the person reading it can see why it is on the list.
The Reddit move in your AI Radar plan now links the exact thread already holding the answer slot, so you can open it and read before you post.
When we email you three days after a scan, the re-scan link now has your page already filled in.
The before and after comparison now agrees with your dashboard about which fixes landed. Reworded findings no longer read as fixed and reintroduced.
Quick wins now rank higher in your fix queue, so the fixes you can finish in minutes sit near the top.

v4.34August 3, 2026

New on the blog: llms.txt, does it actually work? The full data behind the most-searched AEO tactic of 2026, including why we removed the llms.txt bonus from our own scoring engine and what the evidence says to spend that time on instead.
The glossary now defines llms.txt, so you can check the term without reading the whole piece.
The answer engine optimization guide now opens with a direct, quotable answer to the question everyone is asking it: what is answer engine optimization?
How AI decides what to cite now leads with a short answer up top, so you get the mechanism in thirty seconds and the evidence underneath it.

v4.33July 31, 2026

The email that delivers your free AI Radar check is now written in plain English. It leads with what actually happened, splits the two results that matter into separate numbers, and states them as counts you can read at a glance rather than a score out of 100.
Those two numbers are now kept apart properly: how often AI brought you up when buyers described the problem, and how often it named you when the question already contained your name. Every engine and competitor detail underneath is now tied to the right one of the two.
The link in that email now works for a full 30 days, on any device, as many times as you need. That matches how long we keep the report itself, so the link stays good for as long as there is something to open.
Your free check is saved to your email address the moment you run it. Sign in with that address at any time, by any method, and your report opens with the fixes unlocked.
The email now says plainly what the free check covered and what it could not: eight questions we picked for you and ran once. On a paid plan you write the questions yourself, edit ours, and order them the way your pipeline runs.

v4.32July 30, 2026

The AI Radar client report now has a monthly view. A Weekly | Monthly switch on the client link rolls the trailing four weekly scans into one period: a week-by-week trend line with a plain-language read on whether the movement is a real climb, a slide, or noise, plus a month-over-month change on the headline cited rate.
The monthly view surfaces what a single week cannot: which questions sustained a gain or loss across the whole month, naming the competitor and engine that took or lost the answer slots, and linking each mover to the plan action that addresses it.
Competing sources now carry a persistence chip in the monthly view, so an entrenched blocker that held answer slots every week of the month reads differently from a one-week blip.
The monthly engines table shows each engine's change against the month before, and calls out the engine with the biggest gap along with the specific move that engine rewards.
Where an AI engine recommended a rival by name, the monthly report quotes that answer verbatim, so you can see exactly what the work has to displace.
Scan scheduling is simpler: every AI Radar client scans weekly on your chosen day and hour. The weekly trend is the signal; the monthly report is a rollup of those scans, not a slower scan.

v4.31July 29, 2026

Score trend charts now mark the dates the WhyIQ engine itself changed. A small diamond with a dashed vertical line shows each scoring or measurement update, so you can tell a shift caused by our calibration from one caused by your page.
Hover or tap a marker to see what changed, when, and which touchpoints it affects: WhyIQ Score, AI Citability, Search Rank, Accessibility, or AI Radar.
Markers appear on the Scanner score trend, the client overview trend, AI Radar's weekly citation-rate trend, and the Performance report chart. Updates that land close together merge into one marker with a combined note.

v4.30July 29, 2026

AI Radar now has a brand filter across the dashboard: switch any view between All queries, Buyer questions, and Brand searches, so you always know whether brand-name searches are in the numbers you are reading.
On This Week, the filter re-computes the whole screen: the verdict, the per-engine breakdown, the movement feed, and the competitive snapshot. Brand mode flips the headline to your brand-defence rate, showing whether AI engines cite you when people search your name.
In Explore, the Performance report filters every metric card, the weekly chart, all six dimension tables, and the CSV export in one tap, and the Coverage report re-buckets its status bar, trend, and reason tables to the selected query type.
The Prompts page can filter to buyer questions or brand-defense prompts, with a clear count of what is shown.
Every filtered view states its denominator in plain words underneath the toggle, so a screenshot can never pass off a brand-inflated number as buyer-question performance.
The Performance report now shows AI answers checked as its own card beside Citations, Citation rate, and Avg position, so you can see how many attempts sit behind the rate at a glance.
The Performance chart marks weeks where your tracked queries changed. Hover a marker to see exactly which queries were added and which were dropped, so you can tell whether a movement in the trend came from a real shift or from editing your prompt set.

v4.29July 23, 2026

In a full-site audit, every scanned page now opens into its own full page report. Click a page in the page-by-page chart or list to read its complete findings, fixes, visitor psychology, and all four pillar scores, then use the back link to return to the site report.
The per-page report is generated from the site scan at no extra cost. The one part that needs a standalone in-depth page scan (the individual visitor records) is clearly marked, with a one-click link to run one.
This lives on your own view of the report. The white-label client share and the PDF are unchanged.

v4.28July 23, 2026

Your full-site audit and the white-label client audit now read like designed documents instead of spreadsheets. The site verdict is a score gauge with the four pillar scores on one shared scale, so the weakest channel reads at a glance.
The action plan opens with an impact-versus-effort grid that plots every fix and marks the quick wins, so a client can see why fix one comes before fix four.
Pages are shown worst-first as a ranked bar chart, movement since the last audit as a resolved, still-open, and new split plus a per-page score-change chart, and the visitor journey as conversion and exit-reason bars.
The channel appendix now charts the weakest signals for AI Citability, Search Rank, and Accessibility, and shows how many pages each repeated accessibility issue affects.
The white-label client audit draws each page's score history as a trend line, so a client sees the improvement across scans at a glance, alongside a fixed-versus-open progress bar.
Every chart uses your branding accent, prints cleanly in the PDF, and carries through the white-label share link.

v4.27July 22, 2026

The homepage now shows the improvement loop in action: re-scan a page and the dashboard classifies every finding as Fixed, Still open, or Changed, with a score trend line tracking whether your changes worked. This is the real product view, on sample data, and it links straight to a full site audit sample.
AI Radar now has an interactive walkthrough of the actual app on its marketing pages: click through the real This Week view, the Prompts page, and a single-prompt deep-dive (what every one of the 5 engines answered, verbatim), all rendered inside the app on sample data, on the AI Radar overview and for-agencies pages.
AI Radar's marketing now shows the week-over-week trend chart it always described: a real six-week citation-rate line on the AI Radar overview and How It Works pages, so you can see the trend that is the actual product rather than just read about it.
AI Radar's sample report now surfaces the verbatim AI quotes and the honesty methodology up front instead of hiding them below a fold, and the live product view now appears on mobile, not just desktop.
The AI Radar comparison pages against Otterly and Profound now show the white-label weekly client report as a head-to-head row, name the full scan scale (600 AI queries per client every week on Agency), and link to a sample client report.
How It Works now covers all four scores a scan produces, including Search Rank, and adds a step on the fix-and-re-scan loop that a one-shot AI grader cannot give you.
Consistency and clarity pass across the marketing site: WhyIQ simulates 50 visitors across 21 research-calibrated archetypes (stated consistently everywhere), calibration against 200+ peer-reviewed papers, and one clear site-scan page count.

v4.25July 21, 2026

The full-site report is rebuilt as an agency-grade document you can hand a client. It leads with the site verdict, then a short prioritized action plan (each fix stating the problem, the evidence from the simulation, and the change to make), then a page-by-page section ordered worst-first with visitor quotes and, when available, a screenshot of each page.
Site reports now show what changed since the last audit: which issues you resolved, which are still open, and how each page's score moved. The first audit sets the baseline the next one reports against.
PDF export is completely redone. The PDF is now the report itself, printed cleanly, with your branding, a running footer with page numbers, and no leftover website navigation or cookie bars. This covers the site report, the single-page report, and the Executive Brief.
The white-label share link for a site audit now shows the full document your team sees, with your branding, instead of a thinner summary.
Agencies can hide sections, hide individual fixes, override the summary, and add a note, and the share link and PDF reflect those edits.

v4.24July 20, 2026

The free AI Radar check now shows what each engine actually said, quoted verbatim, including the answers where a competitor was recommended instead of you.
Your result now separates buyer-intent citation from brand defence. Holding your own name is a different job from being recommended to someone who has not heard of you, and blending the two was hiding the number that matters.
Every prompt row now shows where you ranked in the answer and which domain took the slot when you were absent.
The headline score carries a typical range, so a small sample reads as a small sample.
Your result page now shows what weekly tracking would add, using your own numbers, with a sample weekly report you can open before deciding.
Your results email now leads with the buyer-intent number and links straight to the sample report and plans.

v4.23July 20, 2026

Agencies can now hand a client one durable link that always shows that client's current audit: what has been fixed, what is still open, and the score movement across every scanned page. Create or revoke it from the client page with Share audit.
The link is white-label. Your logo, agency name, and accent colour carry through, and on the Agency plan the WhyIQ mark can be hidden entirely.
Opening a client now leads with its pages ranked by what needs work first, each opening its own improvement loop.

v4.22July 20, 2026

The Website Scanner dashboard is rebuilt around one question: what do I fix next, and did my last fix work. Your home now lists your pages, with the one that most needs work first.
Open any page to see where it stands, what moved since your last scan (the fixes that worked no longer appear), what is still in the way, and one clear next step.
Page scans and site scans now live together under Scans, with a filter to switch between them.
If you bought a scan pack, you can now open your Pages and Scans. Previously a paid pack could still be treated as a free account there.

v4.21July 20, 2026

If you use both WhyIQ products, Dashboard in the main menu is now a dropdown so you can jump straight to the Website Scanner or AI Radar workspace from anywhere on the site.
Signing in now takes you to the product you subscribe to. AI Radar customers land in AI Radar, Website Scanner customers land in the Scanner, and if you use both you land back in whichever you used last.
The Dashboard dropdown lists your most-used workspace first, so the one you want is always the top item.
AI Radar subscribers can now reach the AI Radar workspace directly. Previously the link could bounce you back to the homepage.
Both product menus stay available once you subscribe, so you can still explore the other product at any time.

v4.20July 19, 2026

The main menu now groups pages by product. Website Scanner and AI Radar each get their own dropdown with an overview, sample report, pricing, agency page, and competitor comparisons, so it is clear which pages belong to which product.
Pricing and For Agencies in the main menu now open small dropdowns pointing at each product's own page, so you land on the plans or agency offering you were actually looking for.
New page: AI Radar for agencies. The full agency story in one place: white-label weekly client reports, the 3-pass confidence band, per-client workspaces, and what it costs per client, with the sample client report one click away.
A new Resources menu collects Docs, Science, the Blog, the AI Citability Playbook, the Glossary, and Our Story in one place.
The comparison hub now supports direct links to each product's section, so the menu can take you straight to Scanner or AI Radar alternatives.

v4.19July 18, 2026

AI Radar redesigned from the ground up around one job: giving you something to put in front of a client every week. Home is now your client list, sorted so the clients that moved this week float to the top.
Every client opens on a plain-language This Week screen: is AI recommending this brand, what changed, who is beating you and where, and what to do next. No jargon on the first screen.
New white-label weekly client report on the Agency tier: a durable share link per client with your branding, a verdict, the week's movement, verbatim excerpts of what the AI actually said, the competing sources, the action plan, and an always-visible note on how the numbers were measured. New weeks publish to the same link automatically; past weeks stay archived with their numbers frozen.
AI Radar now stores verbatim excerpts of the AI answers that cite or mention your brand, so reports can quote exactly what ChatGPT, Perplexity, Claude, Gemini, or Google AI said.
A steady week no longer shows an empty screen: the dashboard surfaces the strongest true forward signal instead, from position improvements inside AI answers to new mentions, a rising mention rate, or an honest note on where the 4-to-8-week off-page clock stands.
The Search Console style views (Performance, Coverage, Competitors) are consolidated into one Explore section per client, one click from the plain-language layer.
See a full sample client report with fictional data at whyiq.ai/rr/demo. Every Radar preview on the marketing site and in the docs now renders the real product components, not a mock.

v4.18July 18, 2026

The weekly AI Radar email now leads with what moved: citations won, citations lost, and queries where AI engines newly mention your brand, each named with its engine and prompt. On a steady week it shows the early signals that move before citations do (brand mention rate, brand-defence hold) and is upfront that off-page work takes 4 to 8 weeks to land.
AI Radar email subject lines now tell you the headline before you open: how many citation wins and losses the week produced.
Every competitor comparison page (Profound, Peec AI, Otterly, AthenaHQ) now carries figures verified against the competitor's own public pricing pages, with the verification date shown on the page. Peec's July 2026 pricing-model change and Profound's current tier structure are reflected.
The For Agencies page now presents the full agency stack: white-label audits from the Scanner alongside weekly AI citation monitoring from AI Radar, with both prices published.
The AI Radar hero now says plainly why the Agency tier's 3-pass confidence band exists: a single run of a probabilistic engine is noise, not a number you can defend in a client report.

v4.17July 17, 2026

New public research page: Landing Page Statistics 2026. Original data from 15,000 simulated visitor sessions across 326 real websites (average scores, why visitors leave, the pricing-visibility problem), plus the key sourced CRO and AI search numbers, free to cite.
Every blog post now ends with a Read next section of related articles, so it is easier to follow a topic across the blog.
The For Agencies page now spells out exactly what a white-label CRO audit from WhyIQ contains, section by section, with a new FAQ entry.
The sample reports now mirror a real report exactly, section for section. What you see before you scan is what you get back.
Site reports now read consistently in the paper theme throughout, including the accessibility findings and the AI citability strategic-moves sections.
On a site report, a page where a pillar was not measured now shows a dash instead of a score. An unmeasured pillar is no longer shown as a zero.

v4.16July 13, 2026

You can now have your scan report emailed to you. While a page scan runs, drop your email and the full report lands in your inbox the moment it completes, so you can close the tab instead of waiting.

v4.15July 11, 2026

You can now apply an unused scan credit to unlock the report you are viewing, straight from the report page. If you hold a credit and the report is locked, a one-click 'Unlock this report' button spends it there and then.
Every unlock link on a report now carries that exact report through checkout, so a $2.99 purchase opens the report you were reading the moment payment completes.

v4.14July 10, 2026

New on the blog: Why Reddit Outranks You on Google (and in AI Answers Too). Reddit's Google visibility grew 1,328 percent in nine months, and it is now the most cited domain in AI answers. The post breaks down what changed, why each AI engine has different taste in sources, and the three moves that still work for a normal website. Every figure is sourced.

v4.13June 30, 2026

The AI Citability Playbook and our answer-engine guides were refreshed against the latest 2026 cross-engine research. Signal weights now match the engine exactly, brand-mention correlation figures are current (r=0.664, Ahrefs 2026), and FAQ guidance leads with visible Q&A content rather than schema markup (Google retired FAQ rich results in 2026).
AI Radar recommendations are now engine-specific: when an engine isn't citing you, the suggested move is the one that engine actually rewards (community and Reddit presence for Perplexity, comparative 'vs' and 'alternatives' content for ChatGPT and Claude, crawlable HTML and a direct first-paragraph answer for Google AI Overviews and Gemini).
Refreshed the recommendation copy across AI Radar (fixes, checklist, and the free check) so review-site and schema guidance reflects current evidence: review listings are useful corroboration rather than a one-click unlock, and structured data is a small-but-positive signal, not the headline lever.

v4.12June 30, 2026

AI Citability FAQ recommendations now lead with making your questions and answers visible in the page content, which is what AI search engines actually extract. Adding FAQPage schema is presented as an optional bonus rather than the headline fix, reflecting that Google retired FAQ rich results in 2026 and the answer content itself is what matters.
AI Citability scoring was recalibrated against the latest 2026 cross-engine research. The signals are now weighted by how strongly each one actually drives AI citations (crawler accessibility and answer clarity weigh more; FAQ schema weighs less), so most pages see a modest, more accurate score.
Refreshed the research behind the AI Citability signals against the latest 2026 cross-engine studies, so the per-engine guidance and recommendations stay accurate and current.

v4.11June 30, 2026

The 'Unclear value prop' finding is now more precise. It is only raised when your headline genuinely doesn't state what you offer, not whenever visitors were confused about something else on the page (pricing, a CTA, layout). General confusion is now reported as its own, accurately-named signal.
Cookie-consent banners are no longer treated like intrusive marketing popups. A legally-required cookie banner is a minor, expected interaction, so it now carries a much lighter weight than a surprise pop-up that interrupts reading.
More reliable headline reading on pages that show a cookie or popup overlay over the hero. When an overlay obscures the top of the page, the report now uses the page's actual headline text instead of a partial reading.
The 'Large form may deter visitors' finding was recalibrated to a gentler, research-aligned weight so a long form is flagged proportionately, not over-penalized.
More precise accessibility-law detection. Section 504 (which applies only to recipients of federal funding) is no longer flagged on private companies just because the page mentions courses or students.

v4.10June 30, 2026

Pricing findings now understand your business model. For bespoke, custom-scoped businesses (agencies, consultancies, custom development, enterprise sales-led software), quote-based or 'contact us' pricing is the norm, so the report no longer treats a missing price list as a major problem. Instead it suggests a lighter touch where it helps: a 'starting from' range and a clearer picture of what happens after someone gets in touch.
For productized businesses where buyers expect to see a price (self-serve software, e-commerce, fixed packages, coaching programs, courses, set-price audits), missing or hidden pricing is still flagged as a real conversion blocker. The report reads the page's own signals to tell the two apart.
The page-context understanding behind every scan was upgraded to a more capable model, so the report reads your business type, audience, and intent more accurately across the board.

v4.09June 30, 2026

The 'Popup interrupted flow' finding now reflects an actual popup or overlay (a cookie modal or email-capture popup that covers your content), instead of being triggered by a countdown timer or bold marketing copy. Genuinely intrusive popups are now caught, and pages that merely use urgency or strong claims are not mislabeled.
A short, punchy headline is no longer flagged as 'Unclear value prop.' A two-word value proposition ('Ship faster') is treated as clear; only a one-word or empty headline counts as too short to state a proposition.
Promotion and offer detection is more accurate. 'Discount spotted' now requires a real promotion (a percentage or money discount, a coupon, a sale) rather than firing on phrases like 'save time'. 'Free shipping' detection respects negation, so 'no free shipping under $50' no longer counts as offering it.

v4.08June 30, 2026

WhyIQ AI Radar now has its own visual identity. The whole Radar product, including its dashboard, marketing pages, share cards, and emails, now uses a distinct teal accent so it is easy to tell apart from the indigo Scanner at a glance. The Scanner is unchanged.

v4.07June 30, 2026

Form findings now read in plain language. 'Form errors with no guidance' is now 'Form doesn't help fix mistakes', with a clear description: the form has no inline help when someone mistypes a field.
That finding is also now treated as a low-impact note rather than a major issue. A form that lacks inline error help is mostly a form-completion and data-quality matter, not a primary reason visitors don't convert, so it no longer weighs heavily on your score or crowds out your top conversion issues. (If you have a genuinely long form, that is flagged separately as 'Large form may deter visitors'.)

v4.06June 30, 2026

A long form is no longer mislabeled as 'Forced account creation.' A page with more than five form fields now reads as 'Large form may deter visitors' (the real issue, with the actual field count), instead of claiming you force visitors to create an account. Those are different problems with different fixes, and we now only describe what we actually detect. A precise forced-account-creation check is on the roadmap.

v4.05June 30, 2026

Cookie-consent banners no longer get mistaken for a sign-up form. Consent dialogs (Cookiebot, OneTrust, and similar) add a row of checkboxes that the scanner was counting as form fields, which could trigger false 'Forced account creation' and 'Form errors with no guidance' flags on pages that have no form at all. Those consent controls, and site-search boxes, are now excluded from every form measurement, so the form issues we surface reflect your real conversion form.
Fixed a visitor-count display where the 'Information overload' stat could read something like '43 of 14'. The number of overwhelmed visitors is now shown against the full set of simulated visitors.

v4.04June 29, 2026

New documentation: a plain-English guide to using WhyIQ. The new Docs hub links two complete product guides, one for the WhyIQ Scanner and one for AI Radar, each with a sticky table of contents and a visual walkthrough. Every feature is explained the same way: what it is, how to use it, and how to read what you get back. Every score, signal, status, and term is defined in plain language, with the steps to use it.
The Scanner guide covers running page, site, and pre-launch HTML scans, the anatomy of your report, the four scores (WhyIQ Score, AI Citability Index, Search Rank, Accessibility) and the eight AI citability signals, how to read a recommendation, the visitor breakdown, your dashboard, sharing and exports, and re-scanning to measure improvement.
The AI Radar guide covers the free check, setting up tracking, the dashboard tab by tab, the five engines, the weekly cadence and email digests, the single-pass and 3-pass confidence band, competitors, and managing prompts. If you know Google Search Console, the guide shows how every Radar number maps to one you already read.
Docs are one click away from anywhere: from the main navigation, the footer, your dashboard, and a quick link under the scores on every report.

v4.03June 26, 2026

New 5-minute guide: Google Just Told You to Stop Buying AEO Hacks. Google published its first official guide to optimizing for AI search, and it is blunt: llms.txt is ignored, chunking is not needed, and rewriting your pages for the machines is wasted effort. The post is an honest 5-minute summary of what Google says to stop doing and the one content bar it says to clear, plus the cross-engine catch the guide leaves out: several of the tactics Google waves off still earn citations on Perplexity, ChatGPT, and Claude.

v4.02June 26, 2026

Form-related fixes are now grounded in the actual fields on your page. The scanner now finds the real conversion fields (name, email, message, and the like) even when a form is collapsed, in a tab, or submitted without a standard form wrapper, so recommendations name the specific field and what it is missing instead of a generic 'add form guidance'. Site-search boxes no longer get counted as conversion fields.
Reports no longer mention a signup step on pages that have no signup. The visitor-exit breakdown only attributes a drop-off to a form when the page actually has a submission form and the visitor's friction is genuinely form-related, and the wording now reflects whether that form is a signup or a contact and enquiry form.

v4.01June 26, 2026

New guide: How AI Search Decides What to Cite. AI engines retrieve and read live pages (a technique Google calls grounding), then fan one question into several sub-queries and cite the sources they trust. The post explains the mechanic with the data, including the tactics Google's own guidance says do not work, and why the only honest way to know if you are cited is to run the real query and read what the engine cited.
AI Radar now makes its method clear: it measures citation the same way the engines produce it. Every week it runs your real buyer questions through ChatGPT, Perplexity, Claude, Gemini, and Google's real AI Mode, and reads back exactly what each engine cited, the same retrieve-and-cite step the engines run to build an answer. All 5 engines are flat from $29/mo with no per-engine add-ons. The page-scan AI Citability Index still predicts readiness from on-page signals; Radar measures the outcome. Two jobs, kept distinct.
The CRO tool now explains how it reads your page: WhyIQ captures your page from the same three inputs an AI browser agent uses, the screenshot, the DOM, and the accessibility tree (the accessibility tree comes through the Chrome DevTools Protocol, the same interface an agent reads it through). So the page our 50 simulated visitors read is the page an AI agent would read. The 50-visitor simulation that runs on top of it is WhyIQ's own method, calibrated against 200-plus peer-reviewed papers.

v4.00June 26, 2026

AI Citability scoring is now aligned with Google's official guidance on optimizing for generative AI search. A present llms.txt file no longer raises your crawler-access score, because Google Search ignores llms.txt and our own research shows it does not move AI citation rates. Your score now reflects only what actually helps: allowing the AI search and training crawlers, serving content in server-rendered HTML so crawlers that cannot run JavaScript can read it, and clean meta-robots directives. We also stopped recommending you build an llms.txt file.
Schema-markup guidance is now platform-specific everywhere it appears. Structured data helps most on Google AI Overviews and rich results, moderately on ChatGPT and Claude, and barely on Perplexity, and it is not required for AI features. The Answer Engine Optimization page and the in-report recommendations now say so, instead of quoting a single across-the-board number.

v3.99June 17, 2026

Pricing is simpler and the entry price is lower. A page scan is now a single $2.99, instead of buying a pack of 5 scans for $10 or 13 for $25. Pay $2.99 to unlock the full report on the page you just ran, and it stays unlocked for good, every prioritized fix, kept permanently. Re-scanning a page after you make changes is a separate $2.99. There is no subscription.
Agency is unchanged at $249/month for 150 scans a month, multi-page site scans, and white-label client reports.

v3.98June 11, 2026

Starting a scan is now one step fewer. Paste your URL on the homepage (or any marketing page) and you go straight to your report, instead of stopping at a 'choose your scan type' prompt first. Single-page scans are the default; agency and SaaS pages still route into the multi-page site scan, and your traffic-source selection carries through either way.

v3.97June 10, 2026

The AI Radar dashboard now reads like Google Search Console, for AI answers. A new Performance report shows Citations, AI answers checked, Citation rate, and Average cite position (the clicks / impressions / CTR / position mapping), with a weekly multi-series chart, toggleable metric cards, date ranges from 4 to 26 weeks, and a compare-to-previous-period overlay.
The Performance report slices by six dimensions, the way Search Console slices by query and page: Prompts, Pages (where your citations land), Engines, buyer-journey Stages, Appearance (how the answer frames you: top pick, listed option, caveated), and Sources (who wins the answer slot when you do not). Click any row to filter, and export any view to CSV.
A new Coverage report answers why each prompt is or is not cited, the way page indexing explains your URLs: every prompt lands in Cited, Mentioned only, Absent (with the competitor holding the slot named), or No AI signal, with a weekly trend and per-status reason tables.
The Overview is now a true overview: one summary card per report (Performance, Coverage, Competitors, Fixes), each opening into the full report. The Prompts page focuses on managing what you submit for weekly checking, like sitemaps, and each prompt's detail page now carries its own weekly trend line.
The AI Radar landing page now shows the new report style: a Search Console concept-mapping section, a rebuilt live demo card cycling through the dimension tabs, and a refreshed sample report. If you use Search Console, you already know how to use this.

v3.96June 10, 2026

Accessibility and AI Citability fix lists now lead with the specific change and why it matters, instead of a projected point gain. The priority, and for accessibility the legal-risk level, still show so you can triage what to do first. We would rather tell you exactly what to fix than guess at a number we cannot measure.
More precise accessibility law detection. A site with no e-commerce, healthcare, or US-jurisdiction signal is no longer flagged under ADA Title III on the strength of a .com domain alone, so the legal-risk read is grounded rather than speculative.
Site-level scores now weight your conversion-critical pages (homepage, pricing, landing pages) more heavily than incidental pages like blog posts, so a weak blog no longer drags your site score down as hard as a weak pricing page. Search Rank scoring also now separates a missing security or crawlability setting from a cosmetic social-share-card gap, so the score reflects real impact.

v3.95June 4, 2026

New Compare hub at /compare lists every WhyIQ head-to-head in one place, plus two new comparisons: WhyIQ vs Microsoft Clarity (the free-heatmap question) and WhyIQ vs VWO (diagnose what to test before you spend the traffic). Each is honest about what WhyIQ does and does not replace.
Two new pages for who you are: For Founders, on validating a landing page before you have any traffic, and For SaaS, on walking your whole funnel with AI citability scored on every page.
A dedicated Conversion Rate Optimization Tool page, plus a Glossary that defines the terms (AEO, GEO, pre-traffic CRO, AI citability, behavioral simulation, the WhyIQ Score and more), each linking to its deeper explainer. Clear definitions the AI search engines can cite directly.
An author page for Ben Little: his photo, what he writes about, and every published article in one place, so readers and AI engines can see who is behind the research.
The blog now has category pages. Filter posts by Conversion Science, Research, AEO and GEO, Psychology, or For Agencies, each on its own page with its own share card.
A friendlier page-not-found that points you to popular pages and a free scan instead of a dead end. Plus broader AI-crawler access and cleaner canonical URLs under the hood, so more answer engines can read and cite the site.

v3.94June 4, 2026

Every remaining page now has its own branded share card. After the last update covered 15 pages, the homepage, all 14 blog posts, and the rest of the marketing pages (about, how it works, for agencies, contact, changelog, privacy, terms, the GEO/AEO guide, and the blog index) now generate a custom paper-brand card with their own headline. Share any link and the preview shows a real card instead of the generic logo.
Guide and explainer pages now show a visible author byline and last-updated date (Answer Engine Optimization, Pre-Traffic CRO, and The Science). Readers, and the AI engines that cite these pages, can see who wrote them and when, rather than having it buried only in the page code.

v3.93June 4, 2026

Comparison and pricing pages now have proper branded share cards. Share a WhyIQ comparison or pricing link on LinkedIn, X, or Slack and the preview shows a custom paper-brand card with the page's own headline and key stat, instead of the generic logo image. Live on all 10 competitor comparison pages plus pricing, AI Radar pricing, the science page, pre-traffic CRO, and the GEO/AEO guide.
AI Radar now describes itself as what it is: an AI search visibility tool for GEO and AEO (generative engine optimization and answer engine optimization). The product page, pricing page, and share previews now lead with AI search visibility across the 5 engines, while keeping the plain-English 'AI citation monitor' description for anyone who has not met the category terms yet.
Pricing reads consistently across every page now: prepaid scan packs from $10 with credits that never expire, plus Agency at $249/mo. A few comparison pages were still describing an older plan structure; they now match the rest of the site.

v3.92June 3, 2026

New guide: Does AEO/GEO replace SEO? It is the question every founder and marketer is asking, and the answer most agencies give depends on what they are selling. We pulled the 2026 data together and wrote the honest version. No, answer engine optimization and generative engine optimization do not replace SEO, they layer on top of it, and the brands winning AI search are the ones already ranking organically. The guide answers the whole cluster (is SEO dead, do I need to do AEO, how do I get cited in AI answers, should I do both) with a named source on every number.
It also explains what actually changed underneath: the authority signal moved from backlinks to brand mentions, each AI engine cites a different set of sources (ChatGPT mirrors Bing, Perplexity leans on community content), and the first 30 percent of a page supplies 44 percent of its AI citations. It ends with the five concrete moves that get a page cited. On the blog now.

v3.91June 1, 2026

The AI Radar free check now reads the homepage you submit before generating prompts. Previously the 8 buyer-intent prompts were inferred from the domain string alone, which could go badly wrong when a brand name pattern-matched the wrong category (whyiq.ai generated prompts about IQ assessment tools instead of CRO). The free check now fetches your page title, meta description, and first heading first, then generates prompts grounded in the actual product category.
Generated prompts now include both non-brand buyer-intent queries (6 of the 8) and brand-defence queries (2 of the 8: brand vs alternatives and brand review). Real buyers search both ways depending on whether they have heard of you, and the free check now reflects that mix.
We now ask for your email when you start the free check, and email your full report to that address the moment the scan finishes. You can close the tab. The email contains a magic link that signs you in and saves the report to your account in one click. No more newsletter-style sign-in gate after the results land.
While your free check runs, the progress page now cycles through the 8 generated prompts under the progress bar so you can see exactly what is being tested across the 5 AI engines. The earlier pulse-dot-and-bar UI gave no indication of what the scan was actually doing.
We now publicly state the free check takes about 10 minutes end-to-end rather than 90 seconds. Quality takes time. We run 8 prompts through 5 AI engines (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews) without cutting corners on accuracy, then email you the moment it lands.
Under the hood, the 5 AI engines per prompt now run concurrently rather than one after another, so the real scan time drops from around 6 to 7 minutes down to the 1 to 3 minute range for most domains. The 10-minute public expectation builds in a buffer for slow or rate-limited engines so the typical scan arriving faster is a positive surprise.

v3.90June 1, 2026

AI Radar checkout opens faster. We replaced the redirect-to-checkout.stripe.com page with an inline embedded checkout that renders right inside our page. Cold-load drops by roughly 1.5 to 3 seconds because the cross-origin redirect and Stripe's own page bootstrap are no longer on the critical path. Same payment form, same security, same domain custom field, no redirect.
Both pricing pages now pre-create your Stripe Checkout session the moment you hover (or keyboard-focus, or tap on mobile) a tier card. By the time you click through, the heavy server round-trip is already done. Saves another 500 to 1500 milliseconds on top of the inline checkout change. Works for the page-scan packs, Agency, and AI Radar SMB and Agency tiers.

v3.89June 1, 2026

The homepage hero now names what you get back when you submit a URL: your WhyIQ Score, prioritized fixes, and a visitor breakdown. The pack price (5 scans for $10, around $2 per scan, credits never expire) sits directly under the CTA so you no longer need to navigate to /pricing to learn what scans cost. A small format hint beneath the URL input clarifies that we need the full URL with https://.
The top trust ribbon now reads '200+ peer-reviewed papers in calibration' instead of the average WhyIQ Score across scanned pages. The avg-score framing was being misread by first-time visitors as the tool's own grade rather than an aggregate across pages we've scanned. The calibration credential is the same authority signal we lead with on /science and the AI Citability Playbook, and it carries the right meaning at a glance on the hero too.
The hero CTA now reads 'Get my free WhyIQ Score' (adding 'free' to match every other surface on the site).

v3.88May 31, 2026

The inline Stripe checkout on the report gate now opens noticeably faster. We pre-connect to Stripe and start loading the secure checkout components and your payment session in the background when the gate first appears on your report, so by the time you click Unlock all fixes the heavy lifting is already done. Cold-load time on a typical connection drops from roughly two to four seconds to under one second.

v3.87May 31, 2026

The AI Radar landing page now leads with what the paid product actually does. Every weekly scan is deep by design: 5 engines, 30 to 40 buyer-intent prompts, 3 passes on Agency, run across hours via async batch so the cited rate is a confidence band instead of single-shot noise. The page now says so directly.
Agency rigor now carries a memorable number. 40 prompts per client x 5 engines x 3 passes = 600 AI queries per client every week. The figure lands on the how-it-works section, the Agency pricing card, the final CTA, and the FAQ, so the depth differentiator vs single-shot competitors is no longer just a claim but a figure you can hold in your head.
When your scan finishes we email you the digest with the engine and competitor that moved this week, and your dashboard stacks every weekly run so the headline number is week-over-week drift, not a one-shot verdict. Both deliverables are now on the SMB and Agency pricing teasers and in the FAQ.
The 'Ninety seconds' tagline on the how-it-works section now reads 'Set it up once. Deep scan every week. Email when it lands.' The 90-second framing stays where it is true (the free hero check). The paid product gets the framing that matches its actual cadence and depth.

v3.86May 31, 2026

Pricing is simpler. Page scans now come as one-time prepaid packs: 5 scans for $10 ($2.00 per scan) or 13 scans for $25 ($1.92 per scan). Pack credits never expire. Buy when you need scans, come back when you need more. No subscription.
Agency at $249 per month is unchanged: 150 page scans per month, multi-page site scans up to 25 pages per run, white-label PDF reports, unlimited client workspaces, 3 team seats. Multi-page site scans are now Agency-only.
Solo, Starter, and Pro monthly subscriptions are retired for new buyers. The new pack ladder is faster to buy, costs less for occasional users, and gives every buyer the full four-pillar scoring (CRO WhyIQ Score, AI Citability, Search Rank, Accessibility) on every scan.

v3.85May 31, 2026

AI Radar Building now scans your homepage automatically when you subscribe and reports your AI Citability score across 8 dimensions (schema coverage, FAQ quality, answer clarity, heading structure, statistical density, content freshness, author attribution, crawler access) plus the top three fixes that lift it most. Re-scan on demand any time (rate-limited to once every 24 hours).
AI Radar Building auto-detects whether your brand already has a Wikipedia article, Wikidata Q-item, Hacker News mentions, Reddit presence, Trustpilot listing, and explicit AI-bot crawler access in your robots.txt. Items you have already shipped are marked done with the evidence link so you do not see 'claim your Wikidata Q-item' copy when you already have one.
AI Radar Building review-site and directory recommendations are now category-aware. B2B SaaS sees G2 / Capterra / GetApp / SoftwareAdvice / AlternativeTo. Ecommerce sees Trustpilot / Sitejabber / BBB / Google Business Profile / Yotpo. Coaching and service businesses see Google Business Profile / Bark / industry-specific directories. Local businesses see Google Business Profile / Yelp / Chamber of Commerce. Consumer products see Amazon / Wirecutter outreach / BBB. The right recommendation for your business type appears; the wrong ones do not.
AI Radar Building now renders a four-phase comprehensive build strategy that combines your homepage audit, detection signals, and business profile into a prioritised plan: Foundation (your homepage), Entity establishment (Wikidata, Crunchbase, LinkedIn), Review-site footprint (category-specific), Content corroboration (HN, Reddit, listicles, podcasts), and Compounding habits. Each move shows priority (critical / important / opportunity), effort (15m / 30m / 1h / 1d / 1w+), auto-detected status, and step-by-step instructions.

v3.84May 31, 2026

Radar is now WhyIQ AI Radar across every customer-visible surface. The product is renamed throughout the marketing site, the /radar app chrome, every email (welcome, scan-complete, site-scan-complete, follow-up, weekly scan), the four competitor comparison pages (/vs-otterly, /vs-profound, /vs-peec, /vs-athena), the three Radar blog posts, the /radar/pricing and /radar/checkout flows, and the llms.txt file that AI search engines ingest. The shorter 'AI Radar' name disambiguates the category for first-time visitors who never met the product before (a 'radar' could be anything; an AI Radar is unambiguously what it does).
The URL stays /radar, your dashboard URL is unchanged, all integrations and subscriptions stay live. The rename only affects what you see; nothing under the hood moves.

v3.83May 30, 2026

New AI Radar competitor comparison page at /vs-athena covering AthenaHQ versus AI Radar on single-shot citation queries versus a 3-pass confidence band, with the same honest framing as the existing AI Radar /vs-X pages.
Lifecycle emails now carry a passive AI Radar cross-mention: a P.S. block in the blog-subscriber welcome email, a sister-product card in scan-complete and site-scan-complete, and a tertiary action line in the 3-day follow-up. Each link tracks its UTM source so we can tell which surface drove the click.
The /llms.txt file now describes AI Radar in full, including the 6 canonical advantages, accurate prompt counts (8 on the free check, 30 weekly on SMB, 40 per client on Agency), and competitor positioning lines against Otterly, Profound and Peec. AI search engines that ingest llms.txt now have a verbose AI Radar description to extract from.

v3.82May 30, 2026

Citation panel and Radar tenant scans now honour the platform-supplied Retry-After header on 429 responses (both standard seconds and OpenAI's retry-after-ms millisecond extension), retry up to 3 times with backoff capped at 5 seconds, and only suspend a platform when the rate-limit is genuinely sustained. Previously a fixed 1500ms sleep was used regardless of what the upstream platform asked for, and only one retry was attempted, so a brief 20ms TPM saturation could fail the whole call when honouring the header would have cleared it.
Internal rate-limit alerts now carry an accurate title and remediation per error category. Previously a rate_limit failure surfaced as 'auth/quota failure' with 'check the API key' guidance, which sent the engineer to the wrong dashboard. Now the alert title names the specific failure class (rate limit exhausted, auth failure, billing issue, permission denied, or quota exhausted) and the remediation text points at the right next action. Rate-limit alerts also dedupe in their own bucket so a throughput problem can't be hidden behind a still-fresh auth alert dedupe window.

v3.81May 30, 2026

Radar Prompts page now lets you hold more prompts than your tier scans weekly and drag them up or down to choose which ones run. A clear cap line shows which prompts fall above (running this week) versus below (held in your bank, not scanned this cycle). Below-line rows render in muted text with a 'Not scanned this cycle' pill so the distinction is unambiguous.
New prompts now land at the bottom of your list by default. They never silently displace a prompt you curated above the cap line. Drag them up when you want them included next cycle, or use the ChevronsUp action button beside Edit and Archive to one-click move to position one.
Free check tenants see the list read-only by default with an explicit 'Reorder' button to opt into drag mode. Paid tiers (SMB and Agency) get drag handles visible by default. Promoting a prompt with low AI presence above the line surfaces a one-time inline warning so you see the trade-off before commit.
Star button retired in favour of the new drag-to-prioritise model. Your sort order now persists on the server. The slot quota line shows a contextual upgrade CTA when you hold more prompts than your tier scans (Free check shows the CTA only after your first scan completes, so the upgrade ask doesn't fire before you experience the product).

v3.80May 30, 2026

The scanner now refuses to run a CRO scan on pages it genuinely cannot evaluate, refunds the credit automatically, and shows a tailored explanation. Three new rejection categories: client-side-only single-page apps that render nothing without JavaScript (we explain that AI search crawlers see the same blank page and suggest enabling SSR or scanning a page that renders without JS); canvas-heavy interactive games or web apps (we explain CRO framing is the wrong tool and suggest scanning the marketing page instead); pages showing an error state like a 500 or 'page not found' on a sparse layout (we explain it's likely a transient server issue and suggest checking the URL in your browser first). Previously, all three cases ran the full 50-visitor simulation and returned a low-value CRO-framed report scoring the blank or error state.
Scans that bounce on one of the three new rejection categories now cost zero credits. The refund happens automatically server-side, like the existing refund on bot-blocked or HTTP 503 failures.

v3.79May 29, 2026

Scan reports for sites that show a cookie consent banner now surface the actual page CTA in the report instead of the cookie Accept button. This previously affected non-English cookie modals (German, French, Italian, Portuguese, Spanish) and English modals on sites where the consent UI appeared without an additional popup overlay. Recommendations downstream of the CTA field now reference the correct page element. Verified on two production sites (osmium.pro, skillviax.com) that previously surfaced 'Alle zulassen' and 'Accept All' as their primary CTA.

v3.78May 29, 2026

Radar pricing now has a full pricing page at /radar/pricing with all 4 tiers, a competitor pricing comparison table, the multi-pass band callout, and refund + cadence terms in one place.
Radar competitor comparisons now live for Otterly, Profound, and Peec at /vs-otterly, /vs-profound, /vs-peec. Each page covers how Radar's 5-engines-flat pricing compares to the funded competitors' per-engine add-ons and where each tool wins on its own terms.
Three new product blog posts on AI citation tracking. Why single-shot tools present noise as signal (the 3-pass confidence band explained). Why weekly tracking beats daily under the 3-month half-life rule. Buyer-intent prompts versus SEO keywords for AI search.
Homepage now introduces Radar as the complement to Scanner in a dedicated cross-sell section so visitors landing on whyiq.ai learn both products exist.
Radar share previews on LinkedIn, Twitter, and Slack now render a Radar-specific card showing the per-engine 3-pass band concept, instead of the generic WhyIQ share card.

v3.77May 28, 2026

Radar now sends a weekly scan email after each tenant's scheduled scan completes. Each email shows two headlines: which AI engine cited you most this week (with the citation count) and which competitor source dominated your non-brand queries (with how many of your queries cited them). One click on the dashboard button takes you to the full Radar breakdown. Sent automatically when your weekly scan finishes; the email goes to your account email and respects your email preferences (toggle in account settings). The WhyIQ founder dogfood tenant continues to receive the internal weekly digest instead.

v3.76May 28, 2026

Radar weekly scans now cover every active prompt on your tenant, not just the ones you have starred. Previously the cron filtered to priority-flagged prompts only, so LLM-generated prompt sets (which default to non-priority) silently underdelivered on the tier promise. The star is now a sort affordance; weekly coverage is automatic for everything in your bank up to the slot quota of your tier.

v3.75May 19, 2026

The Visitor Breakdown section on the page-scan report now explains, for each visitor type, what that group collectively experienced on the page in their own voice. Each archetype card now carries a 1-2 sentence synthesis grounded in that group's actual quotes, confusion points, and exit reasons (for example: 'Discount Danas wanted pricing visible above the fold. Two of three left when the price reveal required a form submit; the third bounced before scrolling'). Previously the section showed bounce rate, understanding, and conversion percentages for each archetype but no narrative explaining what those numbers actually meant for that group.
Each visitor type now reveals the individual simulated visitors that produced its stats. One most-representative visitor card shows by default with full quote, exit reason, emotional state, scan pattern, time on page, and CTA Pull. A 'Show N more' button reveals the rest of that archetype's visitors. Previously the underlying per-visitor data existed in the scan but was never surfaced anywhere in the report; only the aggregated archetype metrics were shown.
Visitor types with only one simulated visitor in the scan now render a clear 'Only one visitor of this type in this scan. See their full record below.' line, with that lone visitor's record expanded by default. Previously single-visitor archetypes received the same metric treatment as multi-visitor archetypes, which overclaimed the certainty of a single-sample finding.
All these improvements appear on the Visitor Breakdown section of the page-scan report and require a paid plan (Solo and up). The section now reads as a narrative explanation of who reacted how, rather than a metric table waiting for interpretation.

v3.74May 17, 2026

The page-scan report now leads with the diagnostic case (your WhyIQ Score, the four pillar scores, page diagnosis, the objections your page does not answer, and how far your visitors scrolled) before the email ask. Anonymous viewers now see the evidence first, then a single email gate, then the prescriptive section. Previously the email gate sat above the diagnostic content, so the value case had to load after the ask.
Free signed-up viewers now see one specific fix unlocked immediately, with the rest shown as compact teaser cards naming the element and impact tier behind a single upgrade ask. The unlocked fix is the highest-impact, most-specific-element finding the engine produced (selected by a deterministic specificity filter, not just the first one in the array). Previously, free users saw every finding with blurred fix sections and no concrete demo of what a fix looks like.
Every upgrade CTA on the report has been rewritten to lead with the outcome and name the price inline. The per-card unlock CTA reads 'Unlock this fix · Solo $19/mo' instead of 'Upgrade to see the fix'; the bottom-of-report CTA reads 'Unlock all N fixes for [your domain] · Solo $19/mo · cancel anytime' instead of 'See plans'; the email-gate Stripe direct-buy reads 'Unlock all fixes · Solo $19/mo' instead of 'Start Solo · $19/mo'; the magic-link email button reads 'Send my link' instead of 'Save my scan' (honest about the immediate next event); 'See all plans' is now 'Compare all plans'.
The 3-day follow-up email now references the specific fix you saw unlocked, names two more locked findings by element and impact, and routes the primary CTA to /pricing (not back to the report). Subject lines lead with the scanned domain ({domain}: 1 of N fixes unlocked, here is what's still locked) so the inbox preview is recognisable. Score-conditional body copy adapts: a strong-score email frames the locked fixes as refinements that move strong pages to excellent, a developing-score email names that a real chunk of visitors are hitting friction the locked fixes diagnose, a needs-attention email names that most visitors are bouncing and even the top two or three fixes usually move the score significantly.
Locked findings on free-tier reports now show element name plus impact dot instead of blurred fix text. This gives a calibrated curiosity signal: you know the engine found something specific on your Hero CTA / Signup form / Pricing page, you just need to upgrade to see the issue and the change to make. Previously the locked-fix block showed only blurred placeholder bars with no element context.
Pricing card sub-line on the email gate's Stripe direct-buy now names the next-tier signpost (for multi-page sites, see Starter) so Solo buyers who realise mid-checkout that they need site scans upgrade instead of refund.

v3.73May 15, 2026

Scan-failure messages are now tailored to the specific HTTP status returned by the site, with separate guidance for site owners versus people scanning a third-party page. A 503 from the origin host now explains that the site is down or rate-limiting, suggests opening the URL in a browser to confirm, and points site owners at their hosting provider (suspended accounts, expired plans, and origin overload are the common causes). 502 / 504 gateway errors and 404 not-found pages get the same site-owner-aware treatment. Previously every non-200 HTTP status returned a single terse 'Page returned HTTP N' message with no next step

v3.72May 14, 2026

Form-related recommendations now name the specific field and propose concrete helper-text copy. Previously, when the scoring engine flagged a form-friction issue, the recommendation read 'add inline error messages' without saying which fields or what the messages should say. The engine now extracts per-field validation diagnostics directly from the page (label, required-ness, helper text presence, error container pattern, inferred validation rule from the field type and pattern attribute), so recommendations now read 'Signup form: Email field has no helper text. Add "Use your work email, e.g. [email protected]" beneath the field.' Concrete examples are generated for email format, phone format, URL format, length constraints, numeric ranges, and required-only fields. Silent high-risk fields (required, no helper, no error container) that no simulated visitor blamed by name also surface so agencies can address them before they become a real-user issue

v3.71May 14, 2026

Refreshed brand assets across every surface. New favicon, new browser-tab icon, and a new link-share preview image (the card that appears when whyiq.ai is shared on LinkedIn, Slack, Twitter, Discord, iMessage). All in the v2.1 paper-cream identity rather than the dark navy from v1.0. Social media banners (LinkedIn, X, YouTube, Facebook) also generated in both light and dark variants from the new asset pipeline. If you have whyiq.ai bookmarked or pinned, the favicon will update next time your browser refreshes its cache

v3.70May 13, 2026

Meta titles tightened across 7 marketing surfaces so they clear Google's SERP truncation cap (~60 chars). Homepage, /pricing, /science, /vs-manual-audit, /pre-traffic-cro, /answer-engine-optimization, and the why-chatgpt-cites blog post all shortened. Google rewrites long titles roughly 2.3x more often than short ones, and rewritten titles lose 5-10% of click-through rate, so this is a measurable CTR lift across the highest-traffic search-landing surfaces
Meta descriptions tightened on the homepage, /pricing, /about, and /science so the load-bearing trust signals (89.5% accuracy, $19-$249 plans, 'free first scan') survive mobile SERP truncation. /pricing description dropped the 'Free' lead in favour of leading with the actual price range, since the n=1 paid customer in the recent 30-day window came through /pricing not the free path
Hero CTA wording is now consistent across the homepage and primary marketing pages: 'Get my free WhyIQ Score'. Documented exceptions stay intentional: /about keeps 'See what WhyIQ finds on your page' (post-narrative emotional state), /pre-traffic-cro keeps the pre-traffic verb, /guides/ai-citability-playbook and /answer-engine-optimization keep the citability-specific verbs since they output a different metric. Removes choice friction for multi-page-journey buyers who would otherwise see 4 different verbs across 4 pages
New 'WhyIQ vs Hotjar (and other heatmap tools)' section + structured data on /how-it-works. The comparison was already on the page in generic 'Heatmap Tools' form; renamed to make the Hotjar brand-defence query explicit and added a WebPage schema node anchored to #vs-hotjar so AI engines (ChatGPT, Perplexity, Claude) extracting 'is WhyIQ a Hotjar alternative' have a clean target on the WhyIQ site rather than third-party speculation
/about hero reorders to lead with the product descriptor and the calibration stats (50 simulated visitors, 21 archetypes, 200+ peer-reviewed papers, 89.5% extraction accuracy) above the founder quote. Agencies and product-team readers see what WhyIQ is before the founder narrative; the quote still anchors the page below it. /about page title now reads 'About WhyIQ: Predictive Behavioral CRO, Built by Ben Little'
/science hero gains a primary scan call-to-action immediately under the engine-description paragraph, so convinced readers have an exit path without having to scroll through 8 sections to find one. The methodology page reads top-of-funnel and bottom-of-funnel at the same time now
/vs-manual-audit currency normalised to USD everywhere (was a mix of £ and $ in the same scent path). Per-scan price updated to $1.66 to reflect the actual lowest-tier cost on Agency
/answer-engine-optimization final-section CTA paragraph now lists what WhyIQ scores (the 8 AEO/GEO signals: crawler access, statistical density, schema completeness, first-paragraph citability, FAQ quality, heading structure, author attribution, content freshness) directly above the scan button. Closes the bridge between the page's AEO/GEO explainer and the citability-specific CTA
AI Citability Playbook trust strip now shows the average AI Citability score across recent scans (not the WhyIQ Score, which is a different pillar). The page sells AI Citability work; the trust signal now matches. Playbook FAQ schema also gains two new topical questions on off-page work (Reddit, listicles, review-site presence) and pre-launch validation, replacing two product-flow questions that were less directly relevant to the AI-citability resource positioning
Homepage HowTo step descriptions expanded from one-line summaries to 35-50 word paragraphs with named research differentiators (89.5% extraction accuracy, 21 archetypes, 200+ peer-reviewed papers). Google AI Overviews ranks HowTo step text directly, so the longer-and-more-specific text is now AI-search-extractable
Internal schema cleanup that's not user-visible but matters for AI search citation: /science Article + FAQ wrapped in a single @graph; /how-it-works gains a WebPage and WhyIQ-vs-Hotjar WebPage node; homepage SoftwareApplication gains screenshot + creator + publisher entity refs; HowTo nodes on /how-it-works, /for-agencies, and homepage all gain proper author + publisher refs to the global #founder + #organization entity graph; PAGE_DATES registry now covers /for-agencies and /pricing so Perplexity reads them as current instead of falling back to the 17-month-old site-published date
Brand-casing preflight guard added that catches lowercase 'whyiq' slipping into JSX prose or schema text. Same pattern as the existing em-dash and token-purity guards. Wired into pnpm preflight and CI. Caught 6 real violations this session (admin page titles, blog body prose, an aria-live announcement) that were emitting the brand in lowercase to screen readers and to AI engines extracting the page

v3.69May 13, 2026

Scan-time messaging is now consistent across the site. Every page that mentions how long a scan takes now uses the same canonical pair: page scans complete in about 2 minutes, multi-page site scans complete in 10 to 15 minutes. Previously /about said '60 seconds', the homepage said '2 minutes', and several other pages said 'under 15 minutes' or '15 minutes per page' (the last is flatly wrong for single-page scans). Search results that quote a duration from one page now line up with what the next page says
The free-tier description now reads identically on every surface that mentions it: 1 free page scan when you sign up, no monthly reset. Previously the homepage FAQ promised 3 free scans per month while the pricing FAQ promised 1 free scan total. Both pages now emit the same structured data to Google and AI search engines, so the snippet you see in search matches what you get on the site
The /vs-manual-audit comparison page is now eligible to appear as a Google rich result with FAQ snippets and Article metadata. Previously the page had no structured data at all, so it could only render as a plain blue-link result in search. The 6 buyer-focused questions on the page (how it compares to a manual CRO audit, scan times, what WhyIQ replaces and does not replace, when to use each, and cost) are now AI-search-extractable

v3.68May 13, 2026

The conversion pillar is now consistently called the WhyIQ Score across every surface: report headers, dashboard tooltips, email templates (page scan complete, site scan complete, 3-day follow-up), pre-traffic CRO landing page, agency white-label callouts, and downloadable site-scan exports. Previously the same number was variously labelled 'Clarity Score', 'Clarity score', or 'CRO clarity score' depending on the surface, which created friction when agencies cross-referenced reports with clients
The agency dashboard's per-client portfolio table and the site-scan page roll-up table now use 'IQ' as the column header for the conversion pillar (was 'CLR'). Matches the four-pillar shorthand used elsewhere on the dashboard (IQ / AIV / RNK / A11Y)

v3.67May 13, 2026

New AI Citability Playbook at /guides/ai-citability-playbook. A research-backed, step-by-step guide to making your site quotable by ChatGPT, Perplexity, Claude, and Google AI Overviews. Covers the 8 signals AI engines weigh (with the weights from the WhyIQ engine), the platform differences across all four major AI search engines, a 5-move 90-day plan with named effect sizes and sources for each move, the myths the published research has actually debunked, and an FAQ with the 8 questions buyers ask before they invest in this work
The site-report 'Site-wide strategic moves' callout (the one that surfaces AI Citability work that lives off-site, like review-site presence and Reddit community work) now links directly to the new playbook instead of the methodology page. CTA reads 'Read the AI Citability Playbook' so the destination matches the action
Footer Product column gains a link to the new playbook between 'AEO and GEO' and 'Pricing'

v3.66May 13, 2026

Site accessibility reports now read coherently when many pages share the same root cause. Previously, when the same five navigation links failed contrast on every page of a six-page scan, the rollup reported 30 separate items. The aggregator now groups failures by selector and WCAG criterion, so the example above collapses to five unique findings each marked 'Fails on 6 pages' and flagged as a systemic issue. Findings that repeat across at least half the pages render in a new 'Site-wide issues' section; findings isolated to particular pages render in a separate 'Page-specific issues' section. Industry-standard tools (axe, Lighthouse, Pa11y) all dedupe this way
The accessibility rollup header reads more honestly. Score and grade now sit in their own block followed by a single line of plain-language framing: 'The score reflects the proportion of elements that pass each WCAG check. The counts below are findings to address.' Critical findings sit in a separate red-tinted card from other findings, and the issue count now leads with 'unique findings' with a follow-up note like 'Across N page occurrences' so the relationship between the score and the count is no longer ambiguous
AI Citability recommendations that require off-site work (build presence on Reddit / YouTube / industry coverage, get cited in third-party comparative listicles) now surface in their own 'Site-wide strategic moves' callout rather than appearing on every page's fix queue. The page-level fix queue stays focused on changes you can ship to the page in front of you; the strategic callout lives on the site-level report and links to the AI Citability methodology
Cross-page numeric inconsistencies in your site copy are now flagged automatically. If two pages disagree on the same stat by more than 2x (for example, one page says a scan takes '15 minutes' and another says '60 seconds'), the report now surfaces a single site-level recommendation pointing at both pages with the exact context. Catches duration claims, accuracy percentages, and user/customer counts. Zero LLM cost because it's a deterministic check, so this runs on every site scan with no impact on scan time
Pages whose meta description advertises content not visible above the fold now get flagged in the Search Ranking dimension. The check compares meta-description tokens against the visible first-screen text (headline, subheadline, primary CTA, above-fold links). When fewer than 30% of meta tokens appear above the fold, the dimension score is capped and the recommendation names the specific tokens that promised something the visible copy doesn't deliver. Visitors arriving from search no longer land on pages whose first screen feels disconnected from the snippet that brought them there
Accessibility-related recommendations now name specific WCAG contrast ratios from the actual measurements rather than inventing them. Previously, when the scoring engine generated a contrast-related fix, it occasionally cited a specific ratio (e.g. '1.29:1') that did not match the lowest measured ratio for the page. The scoring engine now receives every measured contrast failure as structured input and is explicitly instructed to only cite specific ratios from the input list (otherwise use qualitative language like 'low contrast'). Reports now reconcile against axe and Lighthouse when agencies cross-check
Behavioural simulation of B2B SaaS multi-page scans now draws from a wider, more ICP-aligned visitor pool. Pre-fix, the 'qualified visitor' persona (modelling a keyword-aligned high-intent buyer) was only selectable on ecommerce multi-page journeys. B2B SaaS site scans systematically drew from a consumer-skewed mix that under-represented the actual buyer. The constraint is lifted; B2B SaaS site scans should see WhyIQ Scores lift 2 to 5 points on average where the page is actually well-built for its target audience
Individual simulated visitors now know about content that's signposted on the page but not in their reading window. When a page has an FAQ tooltip, an inline 'See pricing' link, or body copy that mentions a category with a 'see our' affordance, agents are told the explainer is signposted nearby and asked to frame their confusion as 'not prominent enough' rather than 'not explained'. The downstream effect is fewer recommendations that tell you to add explanations that already exist on the page
The animated WhyIQ Score demo on the homepage now jumps to its final value when the visitor's browser requests reduced motion, instead of running the full count-up animation. The same change applies to the agency landing page demo. Resolves an edge case where some screen-recording tools and accessibility-mode browsers captured the demo mid-count and read the in-progress number as the page's verdict on the site

v3.65May 13, 2026

Site-scan report downloads now contain everything the on-screen report shows. Previously, when you exported a site scan as Markdown, plain text, YAML, CSV, or the standard PDF, the per-page section was reduced to one confusion, one recommendation, and one exit reason per page. The six tabs of detail visible on the live report (Findings, Fixes, Psychology, Visitors, AI Citability, Accessibility) did not reach the downloaded file. Every site export except the PDF Executive Brief now carries the full per-page depth: all visitor confusions, the full recommendation list with element and issue and fix and impact and effort, the page diagnosis, the exit-reason and archetype breakdowns, every representative visitor quote, scroll depth, below-fold content visibility, the full visitor psychology block (confidence breakers and builders, trust signal effectiveness, processing mode), the per-page AI Citability dimensions with probable AI search queries, the per-page Accessibility dimensions with sample failing elements and applicable laws and user impact scenarios, and the per-page Search Ranking dimensions with the failing checklist and the SERP preview
The standard PDF for site scans is now a comprehensive report rather than a three-page summary. The cover now shows all four pillar scores beside the WhyIQ Score donut. After the executive summary, page-breakdown table, and site-level recommendations, the PDF now contains one or more pages per scanned page (page header with its own score donut and four-pillar mini-scorecard, page diagnosis, every finding and fix, sentiment-coloured visitor voice quote cards, full psychology block with breakers and builders and trust signal tables, full per-page AI Citability section, full per-page Accessibility section with applicable laws, full per-page Search Ranking section with a Google-styled SERP preview). The PDF ends with the site archetype matrix, site affect summary, site exit reasons, three site-pillar dimension pages, and every individual visitor journey trace
The CSV download for site scans is now a multi-section spreadsheet rather than a single flat row per page. Each section starts with a clearly-labelled marker (page_summary, page_confusions, page_recommendations, page_exit_reasons, page_archetypes, page_visitor_voices, page_scroll_depth, page_confidence_breakers, page_confidence_builders, page_trust_signals, page_ai_dimensions, page_ai_recommendations, page_a11y_dimensions, page_a11y_failures, page_a11y_recommendations, page_search_dimensions, page_search_checklist, page_search_recommendations, plus matching site-level sections and the full journey-step trace). Spreadsheet apps render it as one sheet; analysis pipelines can split on the section markers. JSON exports were already complete and are unchanged. The PDF Executive Brief, which is built to be a 2-3 page client-handoff summary, is unchanged

v3.64May 12, 2026

Marketing pages now read cleaner across the homepage, /about, and /pricing. The homepage subtitle no longer carries the literal '15 minutes' time claim that contradicted the '2 min scan' line directly underneath it. The trust strip below the URL input now also states what happens after you click: 'Paste a URL. Get your report in 2 minutes. No account needed.' so the post-click experience is clear before you act
The URL input across the site now uses 'yoursite.com' as the placeholder default instead of 'your-startup.com'. The previous default signalled the tool was only for startup founders, while in practice WhyIQ works for any landing page or website. The agency pages keep their existing 'yourclientsite.com' override since that audience is explicit
The homepage hero's animated WhyIQ Score demo now carries an explicit 'Example scan output, not a live result' caption directly under the score label. Visitors landing for the first time were occasionally reading the animated score as a verdict on their own page rather than as a product demonstration
The /about page now names what WhyIQ does directly under the founder attribution and offers a primary call-to-action path. Previously the page opened with a founder story that never stated the product category above the fold, leaving visitors unsure whether they were reading about a tool, a service, or an agency. The new descriptor reads: 'WhyIQ is an AI landing page and website analyzer. We simulate how visitors experience your pages and tell you exactly why they leave,' and a button below it takes you directly to the scanner
The pricing page now uses one consistent unit on every plan card. Solo, Starter, Pro, and Agency all show 'N page scans per month (or M full site scans)' so the comparison between plans is straightforward. The underlying credit explainer (1 credit per page scan, 3 credits per site scan, mix freely each month) stays below the cards for anyone who wants the full model
A primary call-to-action now stays in view on the pricing page as you scroll through the plan cards. The bar at the top reads 'Not sure which plan? Run a free page scan first' with a 'Get a free scan' button alongside it. The plan-specific buttons on each card sit below the fold on many screens, so the persistent option means you always have a path to start without having to scroll back up

v3.63May 12, 2026

The site report's 'Pages ranked by impact' table now opens each page's full deep-dive inline. Every row in the table is collapsible: click any page and it expands in place to reveal the same Findings / Fixes / Psychology / Visitors / AI Citability / Accessibility tabs that previously lived in a separate 'Page-by-page deep-dive' section further down the report. The standalone section is gone. Each row still shows the same four pillar scores and issue count as before, so the table reads identically until you click in
The four-pillar grid that used to sit above the table (with the 'click a pillar to recolour' instruction) has been removed. The grid was an extra interaction layer on top of information that was already in the table itself; removing it lets the report move straight from the top-line scores into the per-page detail without an interim recolour step
The sample report at /results/site/demo now uses exactly the same layout, sections, and section ordering as a real site scan report. Previously the demo had a bespoke layout that drifted from the live report whenever either side changed. Both pages now share a single underlying component, so anything that ships to the live report is structurally inherited by the sample report on the same deploy

v3.62May 12, 2026

The site-scan loading screen is now a live per-page progress card. Every page you selected appears as a row from the moment the scan starts, with the page you're currently on shown with a left-to-right progress bar and the rest queued below. As each page finishes, its row reveals all four pillar scores (WhyIQ Score, AI Citability, Search Rank, and Accessibility) in the same layout you see on the final report, so you start reading results the second they're ready instead of waiting for the whole scan to complete
A final 'Compiling report' row sits below the page list and lights up once every page has been scored, showing you the engine is working through the cross-page synthesis and visitor-journey simulation that happens after the per-page work is done. When the synthesis lands, the row settles into its complete state and a 'View full report' button appears, giving you a moment of resolution before the screen moves you to your results
The whole screen now uses the paper brand theme: cream card on off-white, indigo accents only, the WhyIQ monogram as the completion mark. The column structure is fixed from frame one, so the card never grows or shifts as scores come in

v3.61May 12, 2026

New research post: 50 visitors, one landing page, three completely different bounces in the first 8 seconds. We ran 50 simulated visitor types through the same SaaS landing page and watched them leave for three completely different reasons in the same window. The post covers the research underneath the 8-second attention span myth, the three elements that decide whether anyone scrolls, and the four moves that win the first 8 seconds for every visitor type without building six separate pages. Live at /blog/what-visitors-see-first-8-seconds

v3.60May 11, 2026

You can now chat with us live from any page on the site. The chat launcher sits in the bottom-right on desktop as a small dark button with an indigo accent ring, and on mobile it shows as a full-width pill at the bottom of the screen with a rotating example question. Tap to open the chat panel and ask anything
Dismissing the chat does not hide it completely. On desktop, dismissing slides the launcher off the right edge of the screen leaving a small handle you can tap to bring it back. On mobile, dismissing sinks the pill into the bottom of the screen with a small grip indicator. Closing or minimising the chat panel from the chat header now automatically parks the launcher into that low-attention state on both viewports

v3.59May 11, 2026

Anonymous scans you ran before signing up now attach to your account more reliably. Previously, if your browser cleared the link cookie between the scan and signup (incognito, ITP browsers, Brave, manual clear, or signup from a different browser profile), the anonymous scan stayed orphaned and disappeared from your history. Your dashboard now also reclaims any unclaimed scans from the last 30 days that came from this device, even if the cookie is gone
The Share and Export menus on every scan report now open downward into the report instead of upward into the page header. The menus previously opened upward and on most screens the format list and share-link options scrolled off the top of the viewport, leaving only a thin sliver of the upgrade CTA visible. Reports now show the full menu the moment you click Share or Export

v3.58May 11, 2026

Search Rank scoring is now substantially more stable on repeat scans of the same page. The Google Snippet Quality dimension inside Search Rank previously produced large swings between scans even when the underlying page had not changed, because the dimension was bucketed into three discrete grades and the score was sensitive to which page element the model attended to. The dimension now produces a continuous 0-100 grade with a calibrated band, so identical inputs produce identical outputs. Validated against a 5-call repeat test on whyiq.ai: max scan-to-scan variance dropped from 45 points to 0 points
The continuous grade also produces more accurate scoring across the corpus. Pages that previously scored at the upper end of the "partial" bucket and pages that previously scored "aligned" with minor real gaps now sit in a narrower middle band that better matches the actual snippet-to-page coherence. If you re-scan a page you scanned in the last month, expect the Search Rank score to move slightly toward the middle of its range

v3.57May 11, 2026

New category guide at /answer-engine-optimization explains how answer engine optimization (AEO) and generative engine optimization (GEO) work and what predicts whether ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews will cite your page. The guide ships with FAQPage structured data so the answers can be lifted directly into AI search results
Two existing blog posts (Why ChatGPT Cites Some Sources and 69% of AI Crawlers Can't Read Your Website) now carry both AEO and GEO terminology in their metadata, so searches using either vocabulary surface the right post

v3.56May 10, 2026

One coherent visual identity across the entire site. The dashboard, scan reports, billing, sign-in, and every other page now share the same paper-and-indigo design language as the marketing pages, instead of switching to a separate dark palette when you log in
Sign-in screen tightened. The redundant wordmark above the form has been removed (the nav already carries it) and the "Email address" label is gone since the placeholder [email protected] already communicates the field's purpose. The form card reads as a clean white surface against cream
/vs-manual-audit comparison page rebuilt in the v2.0 paper theme. The "When to use each" boxes are now visually distinct (indigo-tinted for WhyIQ, neutral for manual audits) instead of bleeding into the page background
Social-share cards (the preview image you see when a scan link is posted to Twitter or LinkedIn) refreshed to the paper palette. New shares show the new look immediately; older cached shares may take time to refresh in social-media caches
Mobile browser URL bar now paints in cream paper, matching the page it sits above instead of fighting it
Floating "back to top" button now appears on every long page on mobile and desktop, not just blog posts

v3.55May 6, 2026

Renamed /methodology to /science and promoted it into the primary navigation. The page is the audit trail behind every score: 200+ peer-reviewed papers, 50 simulated visitors, 21-archetype pool, 5 stacked psychological frameworks, 89.5% extraction accuracy, under 2.1% scan-to-scan variance. Old /methodology URLs 301-redirect to /science automatically
New "How the scoring stays current" section on /science explains how WhyIQ keeps AI Citability Index scoring up to date as platforms (Anthropic, OpenAI, Perplexity, Google) ship changes. Two-tier signal model with annual and quarterly review cadences and a published config-version stamp on every report

v3.54May 6, 2026

Rewrote the /methodology page from scratch as a research-calibration argument and an answer-shaped extraction surface for AI search. Leads with the audit-trail story (200+ peer-reviewed papers, 50 simulated visitors, 21-archetype pool, 5 stacked frameworks, 89.5% extraction accuracy, under 2.1% scan-to-scan variance) and shows the actual parameter table with the paper behind each coefficient
New "WhyIQ vs a one-shot LLM scan" eight-row comparison block covering input depth, visitor count, calibration source, frameworks applied, output structure, validation, variance, and cost basis
New FAQ section with six answer-shaped Q&As and FAQPage structured data, so AI search engines can lift the answers directly when users ask how WhyIQ works, how it differs from ChatGPT, how accurate it is, or how reproducible the scores are
New "What WhyIQ deliberately does not claim" section makes the boundaries of the simulation explicit, so the page earns trust by being honest about what it is not

v3.53May 5, 2026

Improved scan accuracy on ecommerce sites. WhyIQ now adapts its 50-agent simulation to product, cart, category, search, and confirmation page contexts. New ecom-specific visitor archetypes (cart abandoner, gifting buyer, repeat customer, return-risk buyer, shipping-sensitive shopper, mobile thumb browser, price comparison researcher) are added to the pool when scanning a product, cart, or storefront page. Scans surface ecom-specific friction events: surprise shipping cost at checkout, login wall before purchase, variant out-of-stock, review credibility issues, returns policy clarity, and free-shipping threshold visibility
Advanced scan options now offer the 7 ecommerce visitor archetypes. They live behind a collapsed "E-commerce Visitors" section on the persona picker, so the default scan stays clean while ecom merchants can explicitly force which buyer types are tested
SaaS sites with /products/* URLs (Atlassian, Notion, Stripe-style product pages) continue to scan as SaaS — the engine requires both keyword classification AND DOM corroboration before activating ecom mode, so SaaS marketing pages with cart-vocabulary in their copy don't get mis-routed

v3.52May 3, 2026

New blog post: Why ChatGPT Cites Some Sources and Ignores Others (Our 482-Run Data Set). A breakdown of what we found running a weekly citation panel against Perplexity, ChatGPT, Claude, Gemini, and Google AI Overviews, the predictors that actually correlate with AI citation, and the 5-point checklist for building a brand citation moat

v3.51April 30, 2026

Improved the screen shown when an anonymous visitor hits the 1-scan free limit. The status is now an unmissable amber block ("You've used your 1 free scan") instead of a tiny grey label, plan selection is the primary action instead of a 30-day email reminder, and each plan now reads "N scans / month" so the value is unambiguous

v3.50April 30, 2026

Free scan reports now show a direct "Start Solo · $19/mo" option alongside the email signup, so high-intent visitors can skip the email step and start fixing today. Email signup with Google or magic link stays the primary path. The buy option only shows on scans scoring 40 or higher and includes inline checkout, so you never leave the report

v3.49April 30, 2026

New /pre-traffic-cro category page goes deep on what pre-traffic CRO is, when to reach for it, and which failure modes it surfaces. Includes a six-question FAQ block and links to the methodology and the failure-mode taxonomy
Homepage hero subtitle expanded to make the category positioning explicit ("AI landing page analyzer for pre-traffic CRO"), with a sourced stat on why A/B testing breaks under 1,000 monthly visitors
/methodology section 1 now leads with the 89.5% extraction-accuracy validation stat and the "didn't-understand vs didn't-trust" failure-mode distinction
/vs-manual-audit hero rewritten so the AI-landing-page-analyzer-vs-manual-audit comparison is direct and complementary, not adversarial

v3.48April 30, 2026

Marketing pages now serve from a global Cloudflare edge cache. Returning visitors and users in regions with warm cache see the homepage, pricing, and other marketing pages load roughly 25× faster (server response time dropped from ~480ms to ~16ms). The previous CDN layer was routing-only and didn't actually cache; this change adds real edge caching for the first time
The bare-domain redirect (whyiq.ai → www.whyiq.ai) is now handled at Cloudflare's edge instead of round-tripping to the origin server. Every visitor who hits the bare domain saves about 250ms before the page even starts rendering

v3.47April 29, 2026

Marketing pages now meet Google's 'Good' Core Web Vitals threshold across all three measurements (Largest Contentful Paint, Interaction to Next Paint, Cumulative Layout Shift). Pages load and become interactive substantially faster, and Google's search-ranking signal will reward this once the field-data window catches up over the next few weeks
Google Analytics now loads only after your first interaction with the page (a click, scroll, or keypress) instead of in the background while the page is still rendering. Visitors who bounce away with no interaction never download it. Combined with the same treatment on PostHog, the homepage's initial JavaScript payload is dramatically lighter

v3.46April 24, 2026

The homepage now serves from a global edge cache on every visit instead of being rendered on demand. Cold-start render time is gone: the page arrives in tens of milliseconds anywhere in the world
The hero headline, subtitle, CTA, and trust badges now appear immediately and slide gently into place rather than fading in from invisible. Visually almost identical, but the page becomes usable substantially earlier on slower devices
The PostHog analytics SDK no longer loads or parses until your first interaction with the page. Visitors who bounce never download it at all, which keeps initial loads lighter for everyone

v3.45April 24, 2026

Marketing pages now load and become interactive significantly faster on mobile, especially the homepage. The above-the-fold JavaScript bundle is much smaller, the navigation and hero CTA hydrate without blocking the main thread, and the scan-progress page no longer holds extra animation libraries until needed
Auth-aware UI now renders correctly on the very first paint. Logged-in users see the Dashboard link and the right CTA immediately without the brief flash of logged-out state that used to happen after page load

v3.44April 29, 2026

The scan rating widget now appears right after the Findings & Fixes section, where you have just seen the recommendations. Previously it was below the AI Citability Index, Search Ranking, and Accessibility sections at the very bottom of the report, where most people never reached it

v3.43April 29, 2026

Score column tooltips on the Pages, Sites, and client Scans tabs now appear correctly on hover. A layout containment issue was preventing the tooltip popups from rendering above the table
Scan rows on mobile now show a proper summary card. On phones you see the page URL, your WhyIQ score, and the secondary scores (AI Citability, Rank, Accessibility) labelled inline, along with status and date. Previously all columns stacked as unlabelled numbers
The delete button on a scan is now reachable on mobile. It was previously only visible on desktop

v3.42April 29, 2026

The contact page now copies the email address to your clipboard when you click any of the three contact buttons. If your browser doesn't have a mail client configured, the address is already on your clipboard ready to paste.

v3.41April 28, 2026

AI Visibility is now called the AI Citability Index across the whole product. The score measures how citation-ready your page is for ChatGPT, Claude, and Perplexity, and the new name makes that clearer. The 8 underlying dimensions, the score, and the recommendations are unchanged. The honest framing is that this is an index of citation-readiness signals, not a direct measurement of citations in real LLM responses
Your WhyIQ Score now shows its expected re-scan range. Under the score you'll see 'Typical range X–Y' — the band where the score is likely to fall on a repeat scan of the same page. A tighter band means the 50 simulated visitors agreed more strongly. Anonymous viewers continue to see only the score; free-tier and above see the range
Three new diagnostic narratives appear inside your paid report. A messaging coherence note explains how aligned your headline, subheadline, and CTA are. A SERP and page coherence note explains how well your search listing matches your actual page. An AI citability note explains how citation-ready your content is. Each is a single-sentence narrative the engine writes per scan
Benchmark comparison is now uncertainty-aware. Your page only shows 'above benchmark' or 'below benchmark' when the entire confidence range sits clear of the page-type benchmark, instead of flipping label on a single point near the threshold
AI Search traffic source now applies the right visitor weighting. Selecting the AI Search chip on a scan now weights the simulation toward AI-search-verifier and researcher visitor types with appropriate journey depth caps, so the score reflects how AI-search traffic actually behaves
Multi-page journey emotion now carries between pages correctly. When a visitor encounters frustration on one page of a site scan, that frustration influences how they read the next page, the way real visitors behave

v3.40April 28, 2026

Scans now load slow pages reliably. The initial page-load window has been widened to 25 seconds, so heavy WordPress and ecommerce sites loading external fonts, reCAPTCHA, and ad pixels finish before being given up on. Brazilian, EU, and APAC origins benefit the most
When a site is fronted by Cloudflare Bot Management, Datadome, or PerimeterX, the report screen now tells you that's what happened, instead of showing a generic 'Scan failed.' You also get three concrete next steps: scan a staging URL without the protection, ask the site owner to allowlist the WhyIQ scanner, or toggle off Bot Fight Mode if you control the Cloudflare account
Failed scans now show the real reason. Vercel preview deployments with auth protection, slow-loading pages, and bot-blocked sites each get their own tailored heading and explanation, so you don't waste time retrying something that won't work

v3.39April 28, 2026

Pin numbers on the page screenshot now land where each issue actually lives on your page. Positions are now derived from real page coordinates captured during the scan, rather than category estimates. This applies to all new scans going forward
The page screenshot in your report now has a Desktop and Mobile toggle when a mobile layout was captured during the scan. Switch views to see how the page looks at 390px width
The agency dashboard has a new client status table. Every client is ranked by urgency with Critical, Warning, or Healthy status based on their scores and recent movement. The most important accounts are always at the top, with one-click scan and view buttons on each row
The agency fix queue now has a type-ahead search. Start typing a client name, site domain, or page URL to filter the queue without leaving the page. Selecting a suggestion locks the view to that client, site, or page

v3.38April 28, 2026

Visitor outcome labels in reports are sharper. The 'Converted' bucket is now called 'High intent' across the Why Visitors Left section, persona filter chips, and simulated visitor cards. The simulation measures self-reported motivation, not real-world conversion, and the new label reflects what the number actually represents
The High intent bucket now counts only visitors whose motivation clearly crossed the act-now threshold (above 70 out of 100), instead of including the broader still-evaluating range. The numbers are smaller and more meaningful as a result

v3.37April 27, 2026

All four pillar scores (WhyIQ, AI Citability Index, Search Rank, Accessibility) now appear as columns on every dashboard list — clients, page scans, site scans, and the per-client scans tab — so you can spot where any account or page is strongest and weakest without opening it
Hover any score column heading to see what that pillar actually measures. Tooltips cover what WhyIQ, Citability, Rank, and A11y are scoring against
Dashboard typography aligned to the brand system. Headings, subtitles, and section titles now use consistent sizes and weights across every dashboard page (Clients, Pages, Sites, Billing, Settings, client detail). The 'Clarity' label is renamed to 'WhyIQ Score' everywhere it appeared

v3.36April 27, 2026

Site reports now follow the same section rhythm as page reports. The new sticky sub-nav reads Overview, Findings & Fixes, Journey, Pages, AI Citability Index, Search Ranking, Accessibility — so anything you learned navigating a single-page report applies to a full-site report
All your conversion fixes now live in a dedicated Findings & Fixes section near the top of the site report, instead of being tucked under the Overview. Easier to scan, easier to copy across to your team
Pages ranked by impact is more readable. The homepage shows as 'Homepage' instead of '/', every scored page is listed (no more top-12 cutoff), and rows are tighter so you can take in the whole site at a glance
Search Readiness has been renamed to Search Ranking, and now uses the same accent colour across page and site reports. The score and what it measures haven't changed, just the name and visual treatment
Each pillar section in the site report now carries a coloured top stripe matching its sticky-nav swatch, so it's clearer where you are as you scroll
The Overview section is now one consolidated card with internal dividers, instead of four separate cards stacked on top of each other. Less scrolling, same content

v3.35April 26, 2026

The dashboard has been rebuilt around the four pillars (Clarity, AI Citability Index, Search Rank, Accessibility) and a single ranked fix queue. Every tier now leads with what to do next, instead of a wall of scores
New tier-aware dashboard layouts. Free shows your latest scan and biggest opportunity. Solo, Starter, and Pro show your top fixes, recent scans, and (on Pro) a 12-scan trend with weakest page. Agency shows a cross-client fix queue and a portfolio table sorted by urgency
New page report hero: numbered pins on the page screenshot map directly to the numbered fix list, so you can see at a glance where each issue lives on the page
New site report hero: an interactive four-pillar grid recolours the page table by whichever pillar you tap, and pages are now ranked by projected score lift if fixed
New before/after compare view at /results/[scanId]/compare. Run a re-scan after shipping changes and see exactly which issues you fixed and how many points the page moved
Billing page now shows a forecast of where your scan usage is heading, alongside the day count remaining in your cycle and the projected total charge if you stay on the current pace. Overage warnings appear once the forecast is reliable (after day 7 of the cycle)
Sidebar refresh: scan meter footer now shows your tier, scan count, and an upgrade prompt at a glance. The bottom tab bar on mobile is restyled to match

v3.34April 25, 2026

Each simulated visitor now writes in a voice that matches their persona. A Gen-Z TikTok visitor and a 65-year-old retiree no longer sound the same in their quotes and confusion notes — vocabulary, cadence, and sentence structure now vary distinctly across all 21 personas
Excellent pages can now score higher. A new high-intent persona (qualified, keyword-aligned visitor) was added to the pool, lifting the achievable ceiling on genuinely strong pages from ~80-85 into the 88-92 range. Mediocre and poor pages are unaffected
Scan-to-scan variance reduced. Repeat scans of an unchanged page now land within ~3 points of each other (previously up to 7), making it much easier to tell whether a real change moved the score
More accurate scoring for regulated B2B pages. Pages mentioning compliance frameworks (SOC 2, HIPAA, GDPR, PCI, FedRAMP, ISO 27001, etc.) now route to procurement and analytical visitor profiles instead of impulse buyers, producing more relevant findings
Smarter form friction analysis. Form scoring now weights field types by their real friction cost (phone and address fields penalised more heavily than checkboxes and search fields) instead of treating every field equally
Better detection of returning visitors. Personas now distinguish cold returns (weeks later) from warm and same-session returns, with memory and trust effects scaled appropriately
Improved trust signal scoring. Customer logos and testimonials are now scored as separate trust events with calibrated effects, instead of being lumped together with security badges
Visitor insight cards now give each segment a distinct diagnosis. The 'Biggest opportunities' and 'Already converting well' cards previously repeated the same generic line for similar-looking segments. Each card now explains why that specific group bounced or converted based on their motivation profile (price-driven, analytical, founder, instinct-driven, anxious) and their actual bounce, understanding, and conversion metrics
Stronger defence against malicious page content. Scanned pages are now sanitised for prompt-injection patterns before reaching the simulated visitors, so adversarial sites cannot manipulate the scan output by embedding instructions in their copy
Better detection of visual-only urgency cues. Countdown timers and inventory progress bars rendered as images, canvas, or SVG (rather than text) are now picked up via the screenshot analysis and factored into the urgency-tactic scoring

v3.33April 25, 2026

AI Citability Index recommendations now show which AI platforms are most affected by each fix. Each recommendation card displays platform badges (Google AI Overviews, ChatGPT Search, Perplexity, Claude Search) so you know exactly where a change will have the most impact
AI Citability Index reports now include the search queries AI systems most likely receive for your topic. These appear above the recommendation cards and show the buyer-intent phrasing used in real AI searches, so you can see what your content needs to answer
Platform signal data shown in recommendations is sourced from published studies (BrightEdge, Semrush 10M keyword study, Princeton GEO paper, AirOps 548K-page study) and is dated so you know how current it is

v3.32April 23, 2026

The homepage now loads from a global CDN cache instead of a server render on every visit, significantly reducing time to first byte and improving overall Lighthouse scores

v3.31April 23, 2026

Pages now load significantly faster on all devices. Text and headings appear immediately without a mid-read font swap, and the initial page weight is lower so slower connections reach usable content sooner

v3.30April 23, 2026

Report backgrounds are now a warm off-white instead of a cold white, reducing eye strain during extended reading without changing contrast or readability
Individual finding cards, recommendation cards, and visitor cards now sit visually distinct inside their section with a slightly warmer background tone and a subtle shadow, so the two-level structure of the report is immediately clear

v3.29April 23, 2026

Blog posts now appear on a raised surface card that floats over the background, improving contrast and making long articles easier to read
A scroll-to-top button now appears in the bottom-right corner on mobile after scrolling down a blog post, so you can return to the top without hunting for the nav
Text inside cards across the marketing site has been made larger and more readable. Body descriptions, card text, and fix details that were previously too small now scale up properly on tablet and desktop
Audit finding cards in reports and the homepage sample now show larger text for the What, Why, and Fix sections

v3.27April 22, 2026

Report pages now use a light background, making it easier to read detailed audit findings and compare scores
Score color bands are now consistent across every section of the report. Blue means strong (75+), green means good (65-74), amber means needs work (50-64), and red means critical (below 50). Hover any section score to see the full scale
Section headings are larger and bolder so it is clear where each new topic begins as you scroll through the report
AI Citability Index, Accessibility, and Search Ranking dimension bars are now color-coded so you can see at a glance which areas are strong and which need attention
Clicking Overview in the jump navigation returns you to the very top of the report
Jump navigation now includes a Score summary button on the right side that also returns you to the top
Visitor breakdown now surfaces distinct data-driven insights for each audience type instead of repeating the same sentence

v3.26April 21, 2026

Pages now load faster across the board. Images are served in AVIF and WebP formats automatically, which are significantly smaller than PNG and JPEG at the same visual quality. Static assets are cached for longer so repeat visits load almost instantly

v3.25April 21, 2026

Search Ranking section rebuilt from the ground up. Each dimension now has its own card that shows what is wrong, why it matters, and a specific fix in plain language. Fixes reference your actual page title, H1 text, and domain rather than generic instructions
Dimension names updated to plain English: Google Snippet Quality, Social Share Cards, AI Answer Readiness, and Visitor Expectations. Each card has a one-sentence subtitle explaining what it measures and why
Hover over any dimension name for a plain-English explanation of what the score measures
Technical checklist items now appear inside their parent dimension card as a collapsible section instead of a separate standalone bar
Priority and effort labels merged into a single badge per card, for example Quick Win (minutes) or High Impact (hours)
A How to read this guide at the top explains the two-part layout (previews then cards) so you know where to start

v3.24April 21, 2026

Anonymous scan results now prompt you to save your report after one minute of reading. The prompt explains that results disappear when you close the tab and offers Google or email sign-up to keep them permanently. Dismissed once, it does not come back

v3.23April 21, 2026

Report bottom CTA now adapts to your login state. Logged-in users see a re-scan and dashboard prompt instead of a generic sign-up nudge. Anonymous visitors are guided to create a free account in 15 seconds with a direct path to their dashboard

v3.22April 21, 2026

Dashboard onboarding redesigned for new accounts. The overview now shows how many free scans you have remaining instead of immediately pushing upgrade pricing. The scan credits card (which referenced a discontinued product) has been removed entirely
Clients section is now clearly marked as an Agency plan feature with a lock icon, instead of inviting free users to add clients they cannot use
Clicking Scan Now in the navigation while logged in scrolls directly to the scan input bar in your dashboard instead of just reloading the page
First-time scan guidance walks you through each step: entering a URL, picking a traffic source for more accurate simulation, and running the scan. The guidance appears once and does not return after your first scan
The empty dashboard prompt now scrolls you up to the scan bar instead of sending you back to the marketing homepage

v3.21April 21, 2026

AI Citability Index, Search Rank, and Accessibility sections now use the same What-Why-Fix card layout as the main findings. Each issue shows what is wrong, why it matters, and how to fix it in one card instead of splitting scores and recommendations into separate areas
Passing dimensions are summarized in a collapsed green section so you can see at a glance what is working well without it taking over the report
Sections with many issues show the top three by default with a toggle to expand and collapse the rest. Less scrolling when you just want the priorities
Scroll depth chart now explains what Above Fold and First Scroll mean when you hover or tap the labels
Contrast and readability improved across all report sections. Larger indicator dots, clearer text hierarchy between explanations and fixes, and consistent styling from the first finding to the last accessibility check

v3.20April 21, 2026

Clicking "Scan Now" or any "Try Free" link now scrolls directly to the URL input, landing cleanly below the navigation bar on all screen sizes
The URL input field now shows "Enter your website URL" as its prompt text, making it immediately clear what to type

v3.18April 20, 2026

Homepage now shows a live example of your actual scan output: real scores, real recommendations, and the exact card layout you see in your own report. You can see what WhyIQ finds before scanning your page
New improvement tracking showcase: see how LaunchDay raised its score from 39 to 79 over 8 scans with 3 targeted fixes. The interactive chart shows which fixes moved the score and by how much
Hero animation updated: now shows a score trend chart with fix history, then transitions to the four scores from a single scan. Captures both what the product measures and what using it over time looks like
Report cards are sharper across all pages. Contrast raised on recommendation labels, score breakdowns, and secondary score cards. Backgrounds are now solid, making findings easier to read at a glance

v3.17April 19, 2026

Free scans now show the complete report: all findings, all fixes, objections, and scroll depth analysis. No content is hidden behind a signup wall on your first scan

v3.16April 19, 2026

Scanning is now faster for new visitors. Paste a URL, hit Enter, and the scan starts immediately. No extra steps, no menus, no decisions before you see results
Report sections now include context on why each audit matters and how to use the findings, with color-coded navigation so you can jump between CRO, AI Citability Index, Search Ranking, and Accessibility sections at a glance

v3.15April 19, 2026

Scans you run before signing up are now automatically linked to your account when you create one. Your scan appears in your dashboard immediately with the full report unlocked for your tier

v3.14April 19, 2026

Free scan reports now show half your findings upfront so you can see the kind of issues we catch. Create a free account to unlock all findings with specific, actionable fixes for each one
AI Citability Index, Accessibility, and Search Ranking scores now include a preview of your weakest dimension with real data from your page. Upgrade to see all dimensions and recommendations
Upgrade prompts now show a re-scan reminder: make the fixes, re-scan, and track your score improvement over time

v3.13April 19, 2026

All scores now display as clean numbers out of 100 across the entire report. Letter grades (A, B, C, D, F) and text labels (Solid, Excellent, Needs Attention) have been removed for a consistent, easy-to-read scoring system
Tooltips on mobile no longer clip off the edge of the screen. Info icons and score tooltips now reposition automatically to stay within the viewport
Clicking a "Needs attention" item in the overview now scrolls directly to the matching fix in the Findings section with a brief highlight. Previously these links did not navigate anywhere

v3.12April 19, 2026

All tooltips now work on mobile: tap the info icon or chart row to see the explanation, tap anywhere else to close. Previously these required a mouse hover and were invisible to touch users
Edit-mode controls (hide/show toggles, edit text buttons) are now always visible on mobile and tablet instead of requiring hover
Lateral overflow on mobile devices resolved: page content now extends to the full viewport width with no gap on the right edge
Section jump navigation bar now sticks correctly when scrolling on all pages including the demo report

v3.11April 19, 2026

Visitor breakdown card now uses calibrated traffic-light colors for each metric: understood percentage, click intent, and bounce rate each have their own thresholds based on real scan data so the colors mean something
Search Ranking score now appears in the same position and size as the AI Citability Index and Accessibility scores: large number top-left, consistent across all three secondary audit cards
Social share preview now shows a clear confirmation when your og:image is set but cannot be previewed due to browser security rules, instead of a broken image area
Demo report updated to match the current report structure: unified findings list, visitor breakdown insights, and all four score cards in the correct layout

v3.10April 18, 2026

Shared scan links and social cards now show both your CRO WhyIQ Score and AI Citability Index score, with a headline that explains what the numbers mean for your sales funnel
Every marketing page title and description rewritten to match what people actually search for: landing page analyzer, CRO audit tool, pre-traffic testing, and more
Demo report pages are now discoverable by search engines so you can see what a WhyIQ audit looks like before running one
AI search engines (ChatGPT, Perplexity, Claude) can now crawl every public page, not just the homepage and blog

v3.9April 18, 2026

AI Citability Index scoring now prominently featured across the homepage, pricing, and agency pages so you can see its value before running a scan
For agencies: AI Citability Index audits are now highlighted as a client deliverable your competitors aren't offering yet
Scan results: the AI Citability Index Score upgrade prompt now correctly shows Solo as the entry point ($19/mo) and explains the competitive impact of citation gaps
Terminology unified across the product: the feature is now consistently called AI Citability Index scoring everywhere

v3.8April 17, 2026

Plans now start from $19/month (Solo tier) for individual founders who need a few scans per month
Improved AI search visibility: structured data now includes all plan tiers, FAQ answers rewritten for the questions AI engines actually ask, and page headings optimized for search intent
Accessibility improvements: cookie notice now dismissible with Escape key, touch targets enlarged for mobile, and input field contrast improved
Report components now display correctly on all screen sizes down to 320px, with improved tooltip positioning and responsive grids

v3.7April 17, 2026

Choosing a site scan when not on a paid plan now shows an inline explanation with a direct upgrade option and a one-click escape to a free page scan, so no visitor gets stuck
Review prompt now appears correctly for visitors who run a scan without signing in

v3.6April 17, 2026

Homepage input focus on mobile now correctly dims surrounding content and keeps the scan area fully visible

v3.5April 16, 2026

Homepage now shows live scan stats: average page WhyIQ Score and total fixes delivered, updated from real scan data
Cookie notice no longer covers page content on any device or scroll position
Homepage hero simplified: headline leads directly without a category label, reducing visual noise above the fold

v3.3April 15, 2026

See an example scan before committing your own URL: demo link now visible on the homepage hero and sample audit section
Improved scan recovery: if you close your tab mid-scan, the recovery banner now shows which site was scanned and links you back to results
Improved site scan reliability: site scans now complete consistently and results are saved as soon as the scan finishes
Improved trust signal detection: scans now correctly identify featured-on logos, integration logos, and partner badges across more sites
Improved CTA detection: footer links like Contact and Learn more no longer get mistaken for the primary call-to-action

v3.2April 14, 2026

Sign in with Google: connect your Google account as an alternative to magic link. Existing accounts linked automatically if the email matches.
Pages view in dashboard: see all your single-page scans across every client in one place, with scores and dates at a glance (Starter and above)
Scan launcher redesigned with a clearer Page scan / Site scan toggle replacing the previous dropdown

v3.1April 13, 2026

Full-site reports now include a per-page deep dive for every scanned page: findings, recommended fixes, visitor psychology, and archetype breakdown
Journey emotion arc: a visual chart showing how visitor sentiment changes page to page across the site
Site-level AI Citability Index and Accessibility scores now appear in full-site reports alongside the page-by-page breakdown
Site scans now select the 8 most relevant visitor types for your content based on intent, page type, and traffic source, up from 6
Traffic source context now adjusts how deeply simulated visitors explore each page: paid search visitors are more focused, AI/LLM referrals explore more broadly
More pages in site scans now receive full analysis: improved URL pattern recognition means fewer pages are skipped during multi-page runs
CTA detection improved: primary call-to-action buttons are no longer outranked by announcement bars or footer links
Subheadline extraction more reliable on pages with nested or dynamically rendered content

v3.0April 10, 2026

Accessibility & Compliance Score: 7-dimension audit covering WCAG, legal jurisdiction detection, element-level evidence, and fix guidance (Pro and Agency plans)
Accessibility findings backed by the same browser data source Lighthouse uses: computed accessible names, roles, and contrast ratios
Visual accessibility check: scans now identify low-contrast text, small touch targets, and missing visual labels from screenshots
Real-world impact framing: each finding shows how it affects screen readers and keyboard users, not just WCAG clause numbers
Compliance checklists for ADA, EAA, Section 508, UK Equality Act, and 10+ other jurisdictions
Agency workspace: manage multiple clients, each with their own scan history and branding profile
White-label PDF exports: send clients a branded report with your logo, colours, and domain
Report editing: annotate and adjust scan findings before delivering to clients
Shareable result links: share a scan with a client or stakeholder without requiring a login
Full-site scanning: scan an entire site in one run, with automatic page classification and cross-page scoring

v2.1.1April 5, 2026

More accurate scan results on complex pages: dynamic content, nested structures, and JavaScript-rendered pages now parse reliably
Scans no longer fail on pages with loading animations: visual analysis continues even when screenshots take longer than expected
Broader page coverage: more real-world page types now scan successfully, including pages using shadow DOM and embedded iframes

v2.1April 4, 2026

FAQ sections now detected reliably on pages using advanced structured data markup
Visitor reaction counts are now more accurate: each confusion or positive signal is counted once per visitor, not per event
Simulated visitors who don't scroll the full page now account for content they didn't reach, giving more realistic behaviour

v2.0April 2, 2026

V2 psychological simulation engine: dual-process cognition (System 1 / System 2), emotional arc tracking
5 new visitor archetypes: Discount Dana, Senior Susan, TikTok Tara, Worried Walter, Mobile Marco
Emotional journey visualisation in scan results
Pop-up and error message detection: overlays that block conversion are now flagged
Per-field form friction scoring
Emotional responses calibrated to 16 psychological triggers, each with published research values

v1.2March 29, 2026

Magic link authentication (no passwords)
Pro and Agency subscription tiers
Usage metering: 1 free scan, then pay-per-scan or subscribe
Billing portal via Stripe

v1.1March 26, 2026

Mobile visitor simulation with separate mobile viewport rendering
Trust signal detection: reviews, badges, guarantees, countdowns
A/B test platform detection (Optimizely, VWO)
Embedded video content is now read and factored into scan results
Archetype breakdown table in scan results

v1.0March 23, 2026

Initial build: 50-visitor simulation engine
14 core visitor archetypes
WhyIQ Score (0–100), bounce rate estimate, and ranked confusion points
AI diagnosis and actionable recommendations
Representative visitor quotes for each scan

PlanningMarch 21, 2026Where it started

Product concept defined: simulate how real visitors experience a landing page, not just track what they click
14 visitor archetypes researched and profiled across intent, device, psychology, and behaviour patterns
Core architecture decided: browser-level rendering, per-agent LLM reasoning loops, WhyIQ Score as the primary output
Behavioral modifier engine designed: 10 probabilistic modifiers with temporal decay, grounded in UX and decision science research
Full build strategy written: 5 phases, complete file structure, SQLite schema, TypeScript interfaces

Ready to try it?

Paste your URL and get your first audit free.

Run a free scan