Skip to content

More than tracking: fixes and proof

Don't just track what AI says about you. Fix it, and prove it worked.

Buyers now ask ChatGPT, Gemini, Google AI Overviews and AI Mode, Perplexity and Claude which tools to consider. Recognly finds the wrong facts in their answers, drafts the fixes for your team to approve, and proves the change against a holdout. And every week, you see exactly what each engine says.

The free check needs only your website and an email address.

Acme Analytics / Visibility

Appearance rate, worked vs holdout
Consumer answers (UI capture): share of answers from ChatGPT, Gemini, Google AI Overviews and Google AI Mode that mention the brand, weekly
  • Worked questions34%
  • Holdout questions19%
  • 95% confidence band

Appearance rate

34%

Consumer answers panel. 95% CI 28 to 40, 312 answers

Lift vs holdout

+14 pts

Consumer answers panel. 95% CI +6 to +22 pts

Review queue
Community reply for a named expert to post

Disclosure: I lead solutions engineering at Acme Analytics.

For retention across several Postgres clusters, most teams pair pgAudit with a log shipper. Acme keeps 13 months of audit history and exports to S3, if a managed option helps.

  • Disclosure in the post: passed
  • Grounded in 2 approved facts: passed
  • No vote or review requests: passed
  • One client per thread: passed
In the consumer answers panel (UI capture), worked questions went from 18% to 34% (95% CI 28 to 40, 312 answers); holdout questions went from 17% to 19% (298 answers). Answers from Perplexity and Claude are reported in a separate model behaviour panel (API).

Measured across the six AI engines your buyers use

  • ChatGPT
  • Gemini
  • Google AI Overviews
  • Google AI Mode
  • Perplexity
  • Claude

Why teams use it

What AI says about you is costing or winning you buyers today

Recognly gives you the full picture every week and does the work of correcting it, with numbers you can defend.

  • The work is done for you

    Fixes for your site, replies for your named experts and outreach arrive drafted and checked. Your team only approves, and your experts post by hand.

  • Proof, not hype

    Real numbers with 95% confidence intervals, measured against a holdout of questions left untouched, so you can see what your work changed.

  • Wrong facts cost sales today

    An answer with an old price or a missing feature turns buyers away before they reach you. We find those wrong facts and work to get them corrected.

  • See what you cannot see alone

    Every week, what ChatGPT, Gemini, Google AI Overviews, AI Mode, Perplexity and Claude say about you, answer by answer, for the questions your buyers ask.

  • Be ready where buyers are going

    More buyers now ask an AI engine before they visit a website. Being described correctly there is becoming part of being found.

Not just tracking

Tracking tools vs Recognly

Most tools stop at a number. We also do the work of fixing it, and show you what changed.

What you getTracking toolsRecognly
See what AI engines say about your brand
Find the wrong facts in those answers–
Fixes drafted from your approved facts, ready to approve–
Proof of change against a holdout, with confidence intervals–
Scam numbers and lookalike sites flagged–

How it works

From what AI says today to evidence that it changed

A weekly loop your team can run in an afternoon, with a person approving every step that touches the outside world.

  1. Step 1

    Measure

    Track the questions your buyers ask across six AI engines. Every rate comes with a 95% confidence interval and the number of answers behind it.

  2. Step 2

    Find gaps

    See where competitors appear and you do not, which sources engines cite, and where an answer states something wrong about you.

  3. Step 3

    Fix and contribute

    Get drafts for your site and replies for your named experts, grounded in approved facts. A person approves each one, and experts post by hand with disclosure.

  4. Step 4

    Prove lift

    Some questions are held out on purpose and left untouched. Lift is reported against that holdout, so you can tell real change from noise.

Then the cycle repeats: measure, fix, prove, improve. Every week Recognly measures again, shows what your last fixes changed, and finds the next gaps to work on.

Features

Everything the loop needs, in one workspace

Measurement, facts, fixes, review and reporting share one source of truth, so every draft and every number can be traced back.

See all features
  • Answer tracking

    How six AI engines answer your buyer questions, measured on a schedule with confidence intervals.

  • Company Brain

    The approved facts about your company that every draft is checked against.

  • Site audit

    Finds the technical issues that stop AI engines from reading your site correctly.

  • Opportunities

    Public conversations where your buyers ask the questions you track, scored for relevance.

  • Drafts with compliance checks

    Fixes for your site and replies for your named experts, checked before anyone approves them.

  • Reports

    Weekly and baseline reports your team and your leadership can read in five minutes.

  • Alerts

    A heads-up when visibility drops, a new competitor appears or an engine states a wrong fact.

  • Action plan

    One prioritised list of what to do next, built from every signal in the workspace.

  • AI-ready assets

    llms.txt, Organization and FAQ schema and a fact sheet, generated from approved facts only.

  • Industry benchmarks

    How your appearance and citation rates compare with anonymous peers in your industry.

  • CRM attribution

    Leads and pipeline that arrived from AI engines, from HubSpot or Salesforce.

  • Agency portal and white-label

    Run many client brands from one workspace and give each client a branded view.

  • Enterprise SSO

    Sign-in through your identity provider with OpenID Connect.

Our promises

Promises we keep

What you get from the first week. Lite tracks one brand with the weekly report and 5 drafts a month; the higher plans have more.

  • Every week, you'll see exactly what each AI engine says about you.
  • We find every wrong fact and work to get it corrected, and track the time to correct.Lite and above
  • 14-day free trial, cancel anytime.

Honest by design

Numbers you can defend. Engagement you can stand behind.

AI visibility is easy to overstate and easy to game. Recognly is built to do neither, and the rules are enforced in the product, not left to good intentions. Read the responsible engagement policy.

  • Confidence intervals, always

    An appearance rate from 40 answers is not the same as one from 400. Every rate shows its 95% interval and sample size, and small samples are labelled.

  • Lift against a holdout

    AI answers drift on their own. Comparing worked questions with untouched ones shows what your work changed and what would have happened anyway.

  • Real people, own names, clear disclosure

    Contributions come from your official accounts or named employees posting from their own accounts, with their connection stated in the post.

  • No fake accounts, no automated posting

    Recognly has no feature to create or run accounts or personas, never asks for votes or reviews, and never posts to third-party platforms.

  • Judge the result yourself

    No one controls what AI engines say, so we never promise a placement. We fix what they get wrong and measure the change against a holdout, so you can judge the result yourself.

For agencies

Run AI visibility for every client from one place

Manage each client brand in one portfolio, give clients their own view, and deliver reports under your name.

  • Portfolio view with visibility, issues and pending drafts for every brand
  • Client guests who see and comment on their own brands only
  • White-label reports, including on your own verified report domain
  • 25 team seats plus unlimited client guests
See agency pricing

$399 per brand per month, minimum 3 brands.

Client portfolio

  • Acme Analytics

    2 drafts to review

    34%

    95% CI 28 to 40

    312 answers

  • Globex Cloud

    Nothing waiting

    21%

    95% CI 16 to 27

    264 answers

  • Initech HR

    4 drafts to review

    27%

    95% CI 21 to 33

    288 answers

Rates are for the consumer answers panel (UI capture). Clients sign in as guests and see only their own brands.

FAQ

Questions buyers ask us

Something else? Contact us.

What does the product actually measure?

For each buyer question you track, it samples answers from ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, then records whether your brand appears, how early, which competitors are named and which sources are cited. Results are reported as rates with 95% confidence intervals.

How are the answers collected?

Through the AI providers' official APIs and a licensed data provider for consumer interfaces. Answers captured from consumer interfaces and answers from APIs behave differently, so they are always reported as separate panels and never blended into one number.

Do you post on Reddit, LinkedIn or forums for us?

No. The software never posts to a third-party platform. It drafts replies for your named experts, checks them for disclosure and accuracy, and your team approves them. The expert then posts by hand from their own account.

Can you promise we will appear in AI answers?

We promise what we can deliver and prove: every week you see exactly what each AI engine says about you, from the Lite plan up we find wrong facts and work to get them corrected, and every rate comes with its 95% confidence interval. No one controls what AI engines say, so we never promise a placement. We fix what they get wrong and measure the change against a holdout, so you can judge the result yourself. You can try all of it with a 14-day free trial and cancel anytime before your first charge.

Do you need access to our code or website?

No. Connecting GitHub or WordPress is optional. Without it you download the fixes as files and add them yourself. If you do connect GitHub, we only add a draft pull request: we don't read, copy or store your code, and you review and merge it.

What is a holdout?

A set of your buyer questions that is deliberately left alone. Because AI answers change on their own, comparing the questions you worked on with the holdout shows the effect of your work rather than general drift.

What does the free AI visibility check include?

Enter your website and email and we check up to five buyer questions on Claude and Perplexity, and the first question on ChatGPT and Gemini. You get a short report with where you appeared, which competitors were named, which sources were cited and three suggested next steps. It is a small sample, and the report says so.

Does it work for agencies?

Yes. The Agency plan is priced per brand, includes a client portal where clients see only their own brands, and white-label reports on your own domain.

See what AI engines tell your buyers today.

Run a free check on a few buyer questions, or start a 14-day trial on one brand with all six engines.