Colytics AI: The AI Visibility Platform That Shows Its Working
Every number on this site comes out of Colytics. Per-engine share of voice on a frozen prompt set. A page score built from twenty-one checks, each labelled proven, mechanical or correlated. It is in beta, and it runs behind every engagement.
- Runs behind every Attrock engagement
- Every score check labelled by how it was proven
- Beta, direct access by application
This is the platform that produced the numbers below. Attrock’s own site holds 217 citations across AI platforms, spanning 97 distinct pages. They are pulled from the same index Colytics reads, dated, and they move.
A headline client outcome: metric, window, client or category.
What Is the Colytics AI Visibility Platform?
- Where a URL sits on a results page
- One engine: Google
- Keywords as the unit
- Nothing about whether you are quoted
- Whether an engine names you, per engine
- Which competitor it names instead
- Prompts as the unit, frozen and signed
- Whether your page could be quoted at all
What Makes the Score Different
Six Dimensions, Weighted by What Is Known
Evidence and answerability carry half the weight. Page type, structure, trust and freshness carry the rest. The weights follow the strength of the evidence, not a hunch.
Every Check Carries an Evidence Grade
Each of the twenty-one sub-checks is labelled proven, mechanical, correlated or hypothesised. You can see which part of your score rests on hard ground.
Only One Check Is Graded by a Model
Twenty of the twenty-one sub-checks are mechanical. One uses a model to judge answer quality, and it is worth four points of a hundred. The score is repeatable.
The Index Confound Is Disclosed, Not Hidden
Google AI Overviews and AI Mode did not appear in Ahrefs’ citation index before roughly spring 2026, which means any AI-citation growth figure spanning that boundary is partly measuring index coverage rather than performance.
Platform, Tracker and Checker: What Each Word Means Here
A tracker records positions. A checker runs once and tells you where you stand. A platform runs the frozen set monthly, scores every priority page, maps the retrieval set and reports the delta. Colytics is the third.
What This Is Not
Colytics reads citation data from Ahrefs’ AI index and runs its own prompt sets and scoring on top. We say which is which in every report, because the two have different limits.
Who Usually Comes to Us
A result from this service line: metric moved, window, client or category.
Why a Defensible Number Matters Now
Your Buyers Are Already Doing This
94% of business buyers report using AI in their buying process. (Forrester, Buyers’ Journey Survey 2025)
B2B buyers. Sample and geography undisclosed. · C3
And the Click Is Going Anyway
In the first four months of 2026, 68.01% of Google searches ended without a click, up from 60.45% in 2024. (SparkToro / Similarweb, 2026)
US, browser-based. Excludes the Google mobile app. · C7
What an Undefensible Score Costs You
None of this is visible in a single blended number, which is why we built one that decomposes.
What Buyers Ask About Colytics First
"Is It Ready to Buy?"
It is in beta. That means it runs behind every Attrock engagement today and produces every number on this site, and it is not yet a finished self-serve product.
Direct access is by application while it is in beta. Clients get it inside their engagement. The free checker is the public front door.
We will not sell a beta as a finished product, and we say beta everywhere it appears.
"What Does It Measure That Other Tools Do Not?"
We will not name other tools here. What we can say is what Colytics discloses that we have not seen disclosed elsewhere.
Every one of the twenty-one score checks carries a grade: proven, mechanical, correlated or hypothesised. The scoring version is stated on every report. The index confound is disclosed.
Judge any platform on whether it will tell you those three things.
"Where Does the Citation Data Come From?"
Citation counts come from Ahrefs’ AI index. The prompt sets, the per-engine runs, the retrieval-set maps and the page scoring are Colytics.
We say which is which in every report, because the two have different limits. The index has a known coverage boundary. The prompt runs have sampling limits on some engines.
A platform that blends the two into one number is hiding one of them.
"Can I See My Own Brand in It Before Committing?"
Yes. The free AI Visibility Score is a Colytics baseline run once on your domain.
It shows which prompts name you, which name a competitor, and which name nobody. It is the same measurement clients get monthly, run one time with no login.
If the score says you do not need us, it will say so.
"What Does It Not Do?"
It does not write content, earn citations or guarantee anything. It measures and scores.
It cannot see inside editor-embedded assistants like Copilot or Cursor. Those are sampled by prompt rather than indexed, and we label that as sampling rather than dressing it up.
Anyone claiming full coverage of every AI surface is overselling.
How Colytics Runs an Engagement
Freeze the Prompt Set
100 to 150 prompts spanning everyone who touches the decision, given an ID, signed off. The same set runs every month so nobody can move the goalposts.
Run Per Engine, Record Every Source
ChatGPT, Perplexity, Gemini, Google AI Mode and AI Overviews, each run separately. Every domain and page cited is recorded and ranked.
Score Every Priority Page
Twenty-one checks across six dimensions, one score out of a hundred, each check labelled by how it was proven. The scoring version is stamped on the result.
Rank the Fixes
Every failed check becomes a fix with points attached and an effort estimate. Sorted by points, then effort. The list is the work plan.
Re-Run on the Same Version
Same prompts, same engines, same scoring version. Model updates are flagged when the whole category moves together.
What these steps produced: baseline figure, re-measured figure, window.
What Colytics Produces
Per-Engine Share of Voice
Prompt coverage and share of voice for each engine separately, against three named competitors, on the frozen set.
Retrieval-Set Map
Every domain and page each engine cited for your prompts, ranked by frequency. The target list for off-site work.
Page Extractability Score
Six dimensions, twenty-one checks, one number, each check graded proven, mechanical, correlated or hypothesised.
Site-Level Score
Every priority page scored, aggregated to a site view so you can see whether the gap is one template or the whole estate.
Ranked Fix List
Every failed check as a fix, with points and effort. Sorted by points first, effort second.
Monthly Delta, Method Stated
Same version, same set, per engine. Model updates flagged. The index boundary disclosed where it applies.
Colytics vs a Rank Tracker vs a Generic AI Visibility Platform vs by Hand
| Colytics AI | Rank tracker | Generic AI visibility tool | By hand | |
|---|---|---|---|---|
| Measures presence in AI answers per engine | Yes, each in its own column | No | Usually blended | Yes, slowly |
| Frozen, signed prompt set | Yes, with an ID | Keywords only | Sometimes | Yes, if disciplined |
| Page score you can decompose | Yes, 21 graded checks | No | Rarely; usually one number | No |
| Discloses the index boundary | Yes | , | Not that we have seen | Yes, if you know about it |
| Retrieval-set map per prompt | Yes, ranked | No | Sometimes | Yes, very slowly |
| Status | Beta, says so | Mature | Varies | , |
| Where it breaks | Not yet self-serve | Cannot see citations at all | Breaks when you ask why | Breaks at fifty prompts |
We do not name other platforms here. Ask any of them for the three disclosures above and judge the answer.
What We’ve Actually Moved
WhenAvailable’s AI citations grew from 32 to 305 between February and August 2026.
For WhenAvailable we published one blog post a month and earned nine brand mentions a month, every month, for 17 months.
WhenAvailable’s organic traffic grew from 3,020 to 9,127 monthly visits over 17 months, a 202% increase.
WhenAvailable went from 117 to 648 page-one keywords, a 453% increase, over 17 months.
One Platform Result Still in Measurement
Prompt coverage and share of voice, quarter over quarter, against three named competitors.
Page-score delta and citation share for a client run entirely through Colytics, quarter over quarter
A quarter-over-quarter prompt-coverage and share-of-voice result, measured against three named competitors.
Empty on purpose. Better an empty slot than a number we cannot source.
You’re Not Hiring a Freelancer. You’re Hiring a Team.
One person stays accountable end to end: Gaurav Sharma, who founded Attrock.
How Do You Get Access to Colytics AI?
Three Ways In, All Honest About Beta
Inside an engagement, by direct application, or the free one-off checker.
Inside an Engagement
- Full platform, every output
- Frozen prompt set and monthly runs
- Page scores on every priority URL
- Retrieval-set map per engine
- No separate platform cost
Direct Access: Beta
- By application while in beta
- Same outputs, your own team runs it
- Onboarding call to set the prompt set
- Beta terms: features may change
- Priority for existing clients and partners
Free Checker
- One Colytics baseline on your domain
- Which prompts name you and which name rivals
- No login, no credit card
- Emailed after verification
- The front door to the other two
- Number of engines carrying volume in your category
- Size of the prompt set
- Number of priority pages to score
- Number of named competitors
- Markets and languages
- Whether you want direct access or delivery included
A single-market brand with fifty priority pages sits at the light end. A multi-market estate with hundreds of pages and six competitors sits at the heavy end. The free score tells us which you are.
Time from start to first measurable citation gain, and the scope behind it.
Where Colytics Sits in What We Do
Colytics AI Visibility Platform: FAQs
What Is an AI Visibility Platform?
An AI visibility platform measures whether a brand appears in AI-generated answers and which competitors appear instead. It also scores whether the brand’s pages are built to be quoted.
It differs from a rank tracker, which records positions on a results page, because generative engines do not produce a results page.
Colytics is one, in beta, and it runs behind every engagement on this site.
Is Colytics AI Available to Buy?
Not as a standalone licence yet. It is in beta, runs inside every Attrock engagement at no separate cost, and direct access is available by application.
The free AI Visibility Checker is a one-off Colytics baseline with no login.
We say beta everywhere the platform appears, because it means what it says.
How Does the Page Score Work?
Six weighted dimensions: evidence, answerability, page type, structure, trust and freshness. Twenty-one sub-checks underneath, one score out of a hundred.
Each sub-check carries a grade saying how it was proven: proven, mechanical, correlated or hypothesised.
Only one check is graded by a model, worth four points. The rest are mechanical, which is why the score is repeatable.
Where Does the Data Come From?
Citation counts come from Ahrefs’ AI index. The prompt sets, per-engine runs, retrieval-set maps and page scoring are Colytics.
Every report says which number came from which, because the two have different limits.
The index has a known coverage boundary. The prompt runs have sampling limits on some engines. We disclose both.
Which AI Engines Does It Cover?
ChatGPT, Perplexity, Gemini, Google AI Mode and Google AI Overviews, each in its own column.
Editor-embedded assistants are sampled by prompt rather than indexed, and reports label that as sampling.
We do not claim full coverage of every AI surface, because nobody can. The report says which engines carried measurable volume for your category.
Can I Run It on My Own Site First?
Yes. The free AI Visibility Checker runs one Colytics baseline on your domain and emails the result after verification.
It shows which prompts name you, which name a competitor, and which name nobody.
It is the same measurement clients get monthly, run once. If the result says you do not need an engagement, it says so.
How Is Colytics Different From Other AI Visibility Tools?
We will not name other tools. We will say what Colytics discloses that we have not seen disclosed elsewhere.
Every score check carries an evidence grade. The scoring version is stated on every report. The index confound is disclosed where it applies.
Judge any platform, including this one, on whether it will tell you those three things.
What Does Colytics Not Do?
It does not write content, earn citations, or guarantee anything. It measures and scores, and the work is done by people.
It cannot see inside editor-embedded assistants. It does not cover every AI surface and does not claim to.
A platform that claims all of that is the one to be careful of.
Can Agencies Use Colytics Under Their Own Brand?
Reports, yes. Every output exports into a partner’s template with their name on it.
The platform itself stays Colytics-branded while it is in beta.
A branded platform view is on the roadmap without a promised date. Partners who need it now export the data into their own dashboards, which most already do.
See the Platform’s Output on Your Own Domain
The free checker runs one Colytics baseline on your site. It shows which buyer questions already name you, which name a competitor, and which name nobody. No login, no contract, and no call required.
