"Every AI visibility tool hands you a score. Where did the number come from?"
An AI visibility platform built on your own verified search data — not a black-box score.
Ansidex · by Michael D. MarkhamAnsidex tracks whether ChatGPT, Perplexity, and Google's AI Overviews name your brand, cite it as a source, or leave it out — prompt by prompt, engine by engine — alongside the Search Console data that proves what those answers cost you.
ANSIDEXPrompt intelligence · Vantage AI Labs · demo workspace| Prompt tracked | Engine | Answer state | Source |
|---|---|---|---|
| best ai agents for customer support | ChatGPT | Named in answer | vantageailabs.com/agents |
| top llm observability platforms 2026 | Perplexity | Named in answer | vantageailabs.com/blog/observability |
| enterprise rag vendors compared | Google AI | Cited as source | vantageailabs.com/docs |
| ai code review tools worth paying for | ChatGPT | Absent | — competitor named |
The category, defined
What is an AI visibility tool?
An AI visibility tool measures whether AI answer engines — ChatGPT, Perplexity, and Google's AI Overviews — name your brand, cite it as a source, or omit it entirely when buyers ask questions about your category. Verified AI visibility is the stricter form: every measurement traces back to the client's own Search Console data and the named engine response it came from, so the number can be audited rather than trusted.
The distinction matters because most tools in this category report a single proprietary visibility score. A score is a conclusion — it tells you a number moved without telling you which prompt changed, in which engine, or why. When a client asks "what caused that," a score has no answer underneath it.
Ansidex was built the other way around. The prompt list, the engine, the answer state, and the source URL are the product; any summary figure is derived from rows you can open. That constraint is deliberate, and it's the reason the platform exists as a separate thing from the practice that uses it.
"How is this different from the score tools?"
Verified measurement, or a number you have to take on faith.
Both approaches produce a chart. Only one of them can tell you what's underneath it when a client asks.
| Black-box visibility score | Verified AI visibility | |
|---|---|---|
| Where the number comes from | A vendor's modeled index, method undisclosed | Named engine responses plus the client's own Search Console property |
| Can you audit it | No — the scoring model is the product | Yes — every prompt, engine, answer state, and source URL is listed |
| When the number moves | You're told that it moved | You can see which prompt changed, in which engine, and what replaced you |
| Search data behind it | Third-party estimates | Verified property data — impressions and clicks the client can confirm |
| What you hand a client | A score they must believe | Rows they can check |
"Where does the practitioner end and the platform begin?"
The platform encodes the method. The practitioner still does the judgment.
Ansidex is not a side project that happens to sit next to a consultancy. It's the working method of that consultancy, written down in software — the prompt research, the citation ledger, the verification step against real search data, the reporting rhythm. Fifteen years of doing the work decided what the tool measures and, just as importantly, what it refuses to claim.
That's also why it doesn't pretend to be autonomous. The platform tells you what the answers say. Deciding which prompts matter for your market, reading what a competitor's placement actually means, and choosing what to do next is judgment — and judgment is the part that stays human.
You can take the platform on its own, or the platform with the practitioner on top. What you can't do is get a number out of it that nobody can explain.
"What's actually in it?"
What Ansidex measures today.
Every capability listed here is shipped and in use on live client accounts. Nothing on this page is a roadmap item.
Named, cited, or absent — per prompt, per engine
A tracked set of the questions buyers actually ask in your category, checked across ChatGPT, Perplexity, and Google's AI surfaces, recording the answer state and which source the engine drew from.
Search Console as the measurement spine
Rankings, impressions, and clicks come from the client's own connected Search Console property — not a vendor's modeled estimate — including the zero-click keywords AI answers are absorbing.
Brand confirmation and relevance filtering
Brand names get confirmed before they're measured, with aliases where a name is ambiguous, and off-topic prompts are filtered out — so mentions of an unrelated company sharing your name never inflate the numbers.
Clean separation between client accounts
Every client is a separate tenant with its own data boundary — built into the architecture from the start rather than layered on afterwards, which is what makes reselling it viable.
Dashboards and reports under your brand
Agency branding applies across client dashboards and reporting, so the surface your client logs into carries your identity.
Built to carry someone else's name
Agencies are the primary audience, which is the reason the product has a name at all — a person's name can't be white-labeled. Client dashboards and reports carry your brand; Ansidex sits behind them as infrastructure, optionally acknowledged, never in the way.
The partner arrangement — pricing, fulfillment, how much of the work you keep — is a conversation, not a checkout.
POWERED BY ANSIDEXTwo ways in
Who Ansidex is built for.
Agencies
Resell AI visibility under your own brand, with client dashboards and reporting that never mention us. Partner terms are scoped per client.
White-label & partner program →Companies
The platform your team logs into, with a senior practitioner working the plan on top of it — measurement and the work in one engagement.
SEO for SaaS & growth companies →Curious what GEO even is?
The methodology behind the platform: what generative engine optimization is, how citations are earned, and why the click stopped being the metric.
AI SEO & GEO services →"Does the data hold up?"
Measured on live accounts.
Figures from live client accounts, anonymized. No ratings, no testimonials, no case studies invented for this page — when there are real ones, they'll appear here with names attached.
Straight answers
Questions about the platform.
What is Ansidex?
Ansidex is an AI visibility platform that tracks whether AI answer engines name your brand, cite it as a source, or leave it out — per prompt and per engine — and pairs that with verified Search Console data so the measurement can be checked rather than taken on faith. It's built and run by Michael D. Markham, and it encodes the same methodology used in client engagements.
How is it different from an AI visibility score tool?
Most tools in this category report a proprietary score. Ansidex reports the underlying rows: the prompt, the engine, whether you were named, cited, or absent, and which URL the engine drew from. Any summary number is derived from data you can open. If a figure moves, you can see exactly which prompt caused it.
Where does the data come from?
Two places. AI answer states come from monitored responses across ChatGPT, Perplexity, and Google's AI surfaces. Ranking and traffic figures come from the client's own connected Google Search Console property — their data, not a third-party estimate — which is what makes the measurement verifiable.
Can agencies white-label it?
Yes. Client dashboards and reports carry your agency's branding, and the platform is multi-tenant by architecture rather than by workaround, so client accounts stay cleanly separated as you add them. Partner terms are scoped per client on a call.
Can I use the platform without the retainer?
Yes — platform access and the hands-on engagement are separate things. Some clients take the measurement layer alone; others want the practitioner working the plan on top of it. Which fits is part of the demo conversation.
How do I get access?
Through a demo rather than a signup form. Access is set up with a conversation about your category, the prompts that matter in it, and whether the platform is genuinely a fit — which also means you see real output before committing to anything.
See your own answer states.
A demo runs against your actual category — the prompts your buyers ask, the engines answering them, and where you currently stand in those answers. No signup, no card.