Inserviss LabsBe the answer when customers ask AI. And the right one.
See if ChatGPT, Perplexity and Gemini recommend you — and whether what they say is true.
Works with a website, Instagram or just your name
Established brand? Check what AI says about you →Real answers recorded in our published studies — not mock-ups. See the research →
The demand you can’t see
Every dot is a real question customers ask AI. Almost none of them point to you.
An illustrative map for a typical local business. Run the free check and this becomes your real coverage, engine by engine.
What Inserviss does
Get named when customers ask AI.
Then booked and paid, right there.
One engine, six stages, every industry. Below: how it ran on two real engagements — names removed, numbers as measured.
What buyers ask AI — and what AI says back
A critical-infrastructure intelligence company selling data and research to investors, developers and vendors.
Asked about the company by name — each answer checked against its own pages
Why it’s missing or misquoted
Each gap gets a cause — so the fix targets the cause, not the symptom.
- criticalAI crawlers read “0 of 0 projects” on every data page — the real numbers only appear after JavaScript runs, which crawlers don’t do.
- criticalName collision: the brand shares its name with a crypto token and an app. One engine answers about those instead.
- highOne engine rarely retrieves the company’s own site for branded questions — consistent with weak presence in the search index it relies on.
- highStale fund figures — sizes, terms and target returns no current page states — are still quoted.
- medium3 data inconsistencies across pages, e.g. two different figures for the same metric.
Exact changes — for the site they already have
Every page audited for search, answer engines, mobile and speed. Fixes come as copy-paste markup, not advice.
- →Render the real figures in the server HTML, so crawlers read the numbers people see
- →One figure per metric, on every page that states it
- →A one-line disambiguation on the About page and in llms.txt
- →Always the full company name in titles and first sentences
- →Question/answer markup and alt text on the pages that lack them
No claim goes out that the company hasn’t published
Every fact Maker or Anchor uses comes from a ledger built from the company’s own pages. What isn’t published is left out — never guessed.
Off-site: the sources AI already trusts
Engines lean on entity records, indexes and third-party pages. Anchor makes those complete, consistent and correct.
- 1Create the Wikidata entity — from documents the company supplies, nothing else
- 2Verify the site with Bing Webmaster Tools and push key pages via IndexNow — the lever on the engine that missed
- 3Turn each figure the company publishes into a data-card post, approved by its team
- 4Hold anything fund-related for compliance sign-off — enforced, not a convention
Nothing posts without approval. Competitors are never contacted.
Their AI agent does the work — over MCP
The company connects its own coding agent to Inserviss. It pulls ranked tasks, ships the change in its repo, and reports back. Inserviss checks the live page and re-measures.
What buyers ask AI — and what AI says back
A B2B software company whose buyers ask AI which tools to shortlist.
Why it’s missing or misquoted
Each gap gets a cause — so the fix targets the cause, not the symptom.
- highThe site is open to AI crawlers, with llms.txt and a sitemap — but 16 pages never answer a question directly, so an engine has nothing to quote.
- high23 rival tools compete for the same shortlist questions across 5 product lines.
- mediumPages without structured data leave engines guessing what each page offers.
Exact changes — for the site they already have
Every page audited for search, answer engines, mobile and speed. Fixes come as copy-paste markup, not advice.
- →Open each page with a 40–60-word answer to the question it targets
- →Structured data on the pages that have none
- →Question/answer markup where buyers ask in questions
- →Alt text on every image
No claim goes out that the company hasn’t published
Every fact Maker or Anchor uses comes from a ledger built from the company’s own pages. What isn’t published is left out — never guessed.
- ✓Drafts use only facts stated on the company’s own pages
- ✓No invented prices, clients, results or credentials
- ✓Anything unverifiable is left out, not guessed
- ✓A person approves before anything is published
Off-site: the sources AI already trusts
Engines lean on entity records, indexes and third-party pages. Anchor makes those complete, consistent and correct.
- 1Posts on the company’s own LinkedIn and X — each with a tracked link, posted after approval
- 2Wikidata entity kept complete from verified facts only
- 3Review and publisher sites pitched by a person — competitors are never contacted
Nothing posts without approval. Competitors are never contacted.
Their AI agent does the work — over MCP
The company connects its own coding agent to Inserviss. It pulls ranked tasks, ships the change in its repo, and reports back. Inserviss checks the live page and re-measures.
From answer to customer
- 1MeasuredNamed in the AI answerper engine, per question
- 2ObservedVisit from an AI enginereferrer recorded on arrival
- 3ObservedBooking, demo or leadon your page or your own form
- 4InferenceRevenue credited to AIalways labelled, never blended
AI-to-AI
Every service is becoming AI-legible
AI is starting to research, compare, and recommend on a customer’s behalf across every kind of service, not just local businesses. We standardize how a business represents itself to AI today, and are building toward the operating system that lets any service describe itself the same way — that’s the AI Service Infrastructure OS.
Human services
Booked by asking — a doctor, a dentist, a studio, a class. Where Beacon already operates today.
Doctors · Dentists · Beauty · Education · Fitness · Childcare · Home services · Tourism · Hospitality · Personal services
Explore
See what AI already says in your industry
Pick your industry for what our published studies measured — or go straight to the research and the free tools.
Measured in our studies
Free check available
Private Banking & Wealth Management
Wealth clients ask AI who to trust with eight figures, before your team hears about the deal.
Same question, three engines, three different answers.
B2B SaaS & Cybersecurity
Buying committees ask ChatGPT to shortlist vendors before your sales team hears about the deal.
Same question, three engines, three different answers.
Dental Practices
New-patient searches increasingly start — and end — inside an AI answer, not a maps listing.
Same question, three engines, three different answers.
Insurance Carriers
A carrier invisible to AI loses the exact moment a buyer is actively comparing quotes.
Same question, three engines, three different answers.
Venture Capital
LPs ask AI for fund due diligence and get no fund named at all, 64% of the time — the emptiest category we measure.
Same question, three engines, three different answers.
Tutoring Businesses
Parents ask AI for the best tutor nearby. Most tutoring businesses are never named back.
Beauty Studios & Med Spas
“Best lash studio near me” now gets answered by AI before it gets answered by a map.
Yoga & Fitness Studios
AI is already recommending studios by name — most owners have no idea if theirs is one.
Consulting, Accounting & M&A
Boutique practices and enterprise firms alike are being left off the shortlist AI generates.
Luxury & Commercial Real Estate
ChatGPT names about six agents per city. Zillow and Redfin are barely part of the answer.
Legal Services
When AI drafts a shortlist of attorneys for a buyer, most firms never make the list.
Start where you are
What you get — and what happens next
Three ways in, depending on who you are. Same engine underneath — and every number labelled as measured, observed or estimated.
Banks, insurers and regulated brands. AI already names you — the risk is what it says: the rate, the fee, the deadline.
What you get
- Answer-accuracy auditReal customer questions on every engine; every claim checked against what you publish.
- Risk registerWrong numbers, rates, fees and deadlines — ranked by customer and compliance impact.
- Brand-facts ledgerYour published facts as the single source every fix is drafted from.
- Fixes at the sourceOn your pages and the third-party sources engines rely on — then re-checked.
How it runs
- 1Agree the questions your customers actually ask
- 2Every answer checked against your published facts
- 3Risk register and fixes, ranked by impact
- 4Re-checked on every engine, every month
Software, data and infrastructure companies whose buying committees ask AI to build the shortlist.
What you get
- Question universe & rivalsThe questions buyers ask per product line, and who AI names instead — per engine.
- Site audit with exact fixesEvery page checked; copy-paste markup and copy, matched to your platform.
- Your AI agent, connectedOver MCP, your coding agent pulls tasks, ships fixes and reports back.
- Pipeline attributionAI-referred visits and demos — observed where we can see it, labelled inference where we can’t.
How it runs
- 1Map the buyer questions and rivals with your team
- 2Measure every engine; audit every page
- 3Your agent ships the fixes; we verify the live page
- 4Re-measure and report monthly
Dentists, studios, tutors, clinics and agencies — anyone customers ask AI to recommend.
What you get
- Your AI Visibility ScorePer engine — ChatGPT, Perplexity, Gemini — never blended into one number.
- The questions you’re missingWhat customers ask, who AI names instead, and which gaps are winnable.
- A page AI can citeBuilt from what you already have — prices, hours, reviews — with booking on it.
- Re-measured after every fixThe same questions, every engine, 14 days after each change goes live.
How it runs
- 1Free check — about 30 seconds, no signup
- 2Your score per engine and the questions you’re missing
- 3Your citable page goes live, with booking
- 4Re-measured on every engine, then monthly
FAQ
Questions people actually ask
Your move
Find out what AI says about you, before your customers do.
The free check takes about 30 seconds and needs no signup. Established brands can request a full answer-accuracy audit.