# Sayso — full content for AI agents > Sayso (https://ai-sayso.com/) measures whether AI assistants — ChatGPT, Gemini, Claude and Perplexity — name a business when customers ask for a recommendation, reports the share of answers per assistant, and does the work to get the business named. Contact: hello@ai-sayso.com. This file is the full text of the pages of ai-sayso.com in Markdown, generated from the pages themselves. Navigation, forms and the illustrative company names used in on-site demonstrations are left out — those names are invented and are not clients. The short index is at https://ai-sayso.com/llms.txt. --- ## AI Visibility Tracking - AI Brand Visibility Tool URL: https://ai-sayso.com/ AI assistant Asked across ChatGPT, Gemini, Claude and Perplexity You who should I use for accounting in Manchester Assistant AI visibility tracking # AI visibility tracking: does ChatGPT recommend your business? Four assistants answer. You’re in none of them. Sayso is an AI visibility tracking service. We monitor how often ChatGPT, Gemini, Claude and Perplexity name you when your customers ask for a recommendation, and do the work to get you named. Free. You get the answers whether or not you hire us. What you don’t see ## You’re not losing traffic. You’re losing customers you never counted. Someone asks an assistant where to go. It names a handful of businesses. They pick one and never open a search engine. You don’t see a lost visit, a bounce or a dropped call — there was nothing to lose. That is what AI visibility measures: not visits, but whether your name came back at all. - Asked an assistant about what you sell - Saw your name The gap never reaches your analytics. There was no visit to lose. Why now ## Search used to be empty too. That didn’t last twenty years. In the early days of search, getting to the top was mostly a matter of being there before anyone else bothered. Then everyone bothered. Now it takes budget, agencies and years of patience just to hold a position. Assistant answers are at that first stage right now. Almost no one is working on how they get described. The businesses that start while it’s quiet are the ones assistants will have been naming for a year by the time the rest arrive. We can’t tell you how long the quiet part lasts. We can tell you it’s happening. What we actually do ## We don’t stop at the measurement. AI visibility tracking plus the work to get you named. Finding out you’re missing from an answer doesn’t put you in it. Something has to change, and someone has to do the changing. That’s us. Every check turns into work we carry out — on the pages the assistants read, on the directories they trust, on the wording they quote. You see each change logged with a date and a result, and you approve what matters. Nothing to brief, nothing to chase, nothing to learn. Inside your account ## You get your own dashboard. Everything we measure, everything we do, in one place. The AI visibility tracker behind the service, and you can log in whenever you like. Your share in each assistant, the questions where you never come up, who gets named instead, and every change we’ve made with its date and result. Nothing is estimated — open any number and read the answer it came from. Share in each assistant, measured separately Every question, and who was named instead Who gets named instead of you The raw answers, word for word Where the assistants send people, page by page How we measure ## Assistants don’t answer the same way twice. So we measure AI brand visibility as a share, not a yes or no. 01 ### A share, not a yes or no We ask each question several times over, in your city and your language, the way your customers phrase it. What you get is your AI search visibility as a number: how often you were named out of all the times we asked. 02 ### Four assistants, never averaged They disagree with each other, often sharply. We report each one on its own, because a single blended share of voice across AI search engines would describe nobody. 03 ### Small swings aren’t progress Numbers move a little on their own, even when nothing has changed. When the movement is small enough to be noise, we say so instead of dressing it up as a result. See it first ## A full report, open. No form, no call. The same document you get after every measurement, with real numbers in it. Read the whole thing before you decide anything. It carries your share in each assistant, the questions that never bring you up, who was named instead, the pages the assistants read, and the work we carried out that period. The one below is a client's own September report, published with their permission. Pricing ## Two things change the price. Which assistants you want watched, and whether you want the work done as well as the measuring. Assistants --- ## How to Measure AI Visibility - Share of AI Answers URL: https://ai-sayso.com/how-it-works The method # How we measure AI visibility, and what we do about it By Olha Kryva Updated 16 September 2026 Long version, for people who want to know what they are buying before they talk to anyone. Nothing here is a secret: if you wanted to do it yourself, this is what you would have to build. - 6 kinds of question, sorted by what the customer wanted - 4 assistants, each reported on its own - 11 pages behind a single answer, on average 01 ## Where the questions come from A customer looking for a supplier does not type your name. They describe the situation they are in: what they need done, where they are, how soon, and how much they want to spend. Those are the questions we measure. We write them from your services and your city, in the language your customers actually use — which in many cities means more than one. Then we sort them by what the person wanted: a service, a price, someone nearby, someone they can trust, or help right now. That split is what turns a single visibility figure into something you can act on. Questions that name a company are kept apart and marked as brand questions. Someone who already knows your name is not choosing any more, so those answers tell you nothing about whether you get found. How the questions are sorted - General who should I use for this in my city - Service a named job: implants, payroll, a leaking roof - Price how much, and where it is cheaper - Local near me, in this district, on this street - Trust reviews, guarantees, who to believe - Urgent today, tonight, this week 02 ## Why we ask the same question more than once Assistants do not answer identically twice. Ask the same question in the morning and in the afternoon and the list can change, even though nothing about your business has. So we ask each question several times over and report a share — how often you were named — instead of a yes or a no. A share can be compared between measurements; a single answer cannot. It also gives us a sense of the noise. Between two measurements with nothing changed, the figure drifts on its own. That is why a small movement we report as no change rather than dress it up as progress. Two measurements, nothing changed in between First measurement 46% Second measurement 49% 03 ## Why we never average the assistants ChatGPT, Gemini, Claude and Perplexity read different sources and answer differently. In our own measurement, four assistants asked the same questions named 63 businesses between them — and only four of those were named by all four. Forty-six appeared in one assistant and in none of the others. An average across four such numbers describes nobody. It hides the assistant you are absent from, which is exactly the one worth working on. Every figure we report belongs to one assistant. 63 businesses named across four assistants - 4 named by all four - 13 named by two or three - 46 named by one, and by none of the others 04 ## What an answer is built from An answer is assembled, not quoted. The assistant reads what it can find about your market — your own pages among them, and alongside them directories, maps, review sites and round-up articles with titles like “best accountants in Manchester” — and writes a short list out of all of it. In the answers we measured, a single answer rested on eleven pages on average. Your own site is one of them, and it is the one source you control completely: it is where an assistant checks what you actually do, for whom, in which city, and at what price. So the work runs on both sides at once. Your pages are made to answer the questions people ask, in plain words a machine can read without guessing. The pages other people keep about you are corrected so that they agree. Where the two disagree, the assistant usually believes the others — which is exactly why both halves are the job. What one answer was built from - Your own pages services, prices, hours, the plain facts about the company - Directories and maps where your details are repeated, right or wrong - Review sites what customers said, and how recently - Round-ups “best accountants in Manchester” and its like - Local media anything written about the market you work in 05 ## What we change Your details, made identical everywhere. Name, address, hours, services and phone number differ across directories more often than not, and an assistant reading three versions of you will often use none of them. Presence in the lists that get read. We find the round-ups and directories the assistants actually cite for your questions, and get you into the ones you are missing from — by the ordinary route: contacting whoever maintains them, with the correct information. The pages on your own site that answer the questions people ask, marked up so a machine can read them without guessing, plus a page of plain facts about the company for the same reason. What we do not do is invent reviews, generate filler articles, or place your name on pages built to be read by machines. Assistants are already learning to discount that, and you would be paying for a problem. Three kinds of work - Your details, identical everywhere Name, address, hours and services differ across directories more often than not. - Into the lists that get read The round-ups and directories the assistants actually cite for your questions. - Pages and markup on your own site The questions people ask, answered plainly, in a form a machine can read. 06 ## What you get, and when The dashboard opens with the first measurement: every question, every answer word for word, who was named instead of you, and the pages those answers came from. Nothing is summarised away and nothing is estimated. After each measurement there is a report — the same document, page for page, that you can read in full on the sample report page before you decide anything. Next to it the work log: what we changed, on what date, and what happened after. First movement usually shows after six to eight weeks, because the pages the assistants read have to change first, and then be read again. If nothing moves, that will be in the numbers too, and we will say so rather than dress up a small swing as progress. What lands, and when - Day one the dashboard, with every answer we collected - After each measurement a report you can send on, and the work log with dates - Six to eight weeks the first measurement where a change could show Read a whole report → ## See where you stand The free check takes one email and tells you whether the assistants name you at all. --- ## AI Visibility Tracking Pricing - Plans and Cost URL: https://ai-sayso.com/pricing Pricing # AI visibility tracking pricing: from a free check to continuous work Two things change the price: which assistants you want tracked, and whether we do the work or you do. Tell us your market and the figure comes back the same day, by email — there is no call to sit through. Assistants tracked ChatGPT The largest audience by a wide margin. Costly to run, and the one most businesses cannot afford to skip. Gemini Answers inside Google’s own surfaces, and cheap to measure. Claude Smaller audience, heavily business and professional. Perplexity Search-shaped, quotes its sources openly. ChatGPT and Claude cost several times more to query than Gemini and Perplexity. That is a real difference in our costs, and it is the reason the price moves. Tracking only the cheap ones is watching the wrong place. Implementation We do the work On by default. Switch it off and you get the measurement, the report and a written plan in order of priority — your team does the work, and the price drops. ## Express check Free one-off Enough to see whether you have a problem. - One assistant of your choice - Your share, and the names that came back instead - Enough to know whether it is worth going further ## Full scan from $150 one-off The whole picture, and what we would change first. - The assistants you choose, measured in depth - Gaps by question type, and by assistant - Every raw answer and the pages behind it - Dashboard access from day one - A written list of what to fix, in order ## Foundation from $300 one-off The scan, the work, and a second measurement to check it. - Everything in Full scan - The changes carried out for you, published and logged with dates - Nothing goes live without your approval - A second measurement eight weeks later to check it moved ## Monitoring from $250 per month For businesses that want the number to keep moving. - A measurement every two weeks - Dashboard with the full history - Ongoing work every month, not just reporting - Nothing goes live without your approval - An alert when your share moves enough to matter No minimum term on Monitoring. Stop whenever you like. ## Paying ahead Monitoring is billed monthly with no minimum term. Pay for a longer period up front and it costs less, because it costs us less to run. - 5% off 3 months up front - 10% off 6 months up front - 15% off 12 months up front - No minimum term. Stop at the end of any month. --- ## AI Visibility Report - AI Brand Visibility Sample URL: https://ai-sayso.com/sample-report Sample report # AI visibility report: a full sample, open A whole report, page by page, before you pay for anything. This is the September report for LFG Hub — a real client, published with their permission. Nothing is invented, nothing is taken out: the numbers, the questions they lose and the plan are the ones they received. Eight pages, the file itself 1 The cover: the four assistants, each with its own share, and the period measured 2 The summary — what the measurement says, and the findings it rests on 3 The questions where the assistants name somebody else, and who that is 4 How the assistants describe this client, in their own words 5 The opportunity map: how crowded each topic is, and whether they are named there 6 Where the assistants send people, then the work that follows from the findings 7 The rest of the plan — each item names the finding it grew from 8 The closing page: who prepared it, when it was measured, what comes next ## See your own numbers The free check is one email and no call. You keep the answers whether or not you hire us. --- ## Free AI Visibility Checker - Is ChatGPT Naming You? URL: https://ai-sayso.com/free-check Free check # Free AI visibility checker An AI search visibility checker with a person behind it. Tell us what you do and where; we ask ChatGPT, Gemini, Claude or Perplexity the questions your customers ask, and send you what came back — the answers themselves, not a score out of ten. - A free AI visibility tool with a person behind it, and no call to sit through. - You keep the answers whether or not you hire us. - We don’t need access to your website. - Usually back within 30 minutes - No card, no account - Raw answers included --- ## AI Search Optimization Services - GEO and AEO URL: https://ai-sayso.com/ai-search-optimization-services Done for you, not sold as a seat # AI search optimization services that get you named Generative engine optimization done as a service, not sold as a dashboard. We measure whether ChatGPT, Gemini, Claude and Perplexity name your business, fix the sources those answers are built from, then measure again to show what moved. - Four assistants, never averaged - Every figure traceable to a raw answer - No minimum term Assistant answer — example Who should I use for an emergency appointment this week? Three places are worth trying. Harbour Street Practice takes same-day bookings, your business publishes its urgent-care prices, and Northgate Clinic has evening hours. - your site - directory - review site - local roundup - trade listing The work is getting your name into that middle clause. It starts with the pages of yours those sources read. - 63 businesses named across four assistants asked the same questions in one local market - 4 of them were named by all four. The rest depend on which assistant your customer happens to open - 11 pages behind an average answer. Your own are the ones that set the facts the rest repeat - 0% share on price questions for a business sitting in the search top three. Visible and invisible at once See a real report, open in full A client's own report, published with their permission — every figure above comes from measurements like it. Read the study behind these numbers Method, sample, limitations and the figures recomputed from the stored answers. What the service includes ## Six things an AI search visibility service has to cover A tool reports that you are absent. A service finds out why and does something about it. All six are included at every tier above the free check. ### The questions your buyers actually ask Written from your site, your services and your city — not from a keyword export. People type keywords into search boxes and whole sentences into assistants, and the phrasing decides who gets named. ### A share you can check, per assistant Every question asked several times on each assistant, because the answers differ between identical asks. One check is an anecdote; five is a number. ### Your own pages, written to be quotable This is where most of an engagement goes. Retrieval lifts passages that stand alone, so service pages answer the buying question in the first two sentences, prices and areas are stated plainly, and the structure is something a machine can parse without guessing. ### The places that repeat what your site says Directories, review platforms and roundups are read alongside your pages, and a contradiction between them is a reason to prefer somebody else. We make them agree with your site rather than the other way round. ### The changes carried out, not recommended We do the work ourselves — listings corrected, pages rewritten, facts made to agree across sources. Nothing is published without your approval, and everything is logged with a date. ### A second measurement that checks it moved The same questions, the same assistants, the same method. Movement under about five points is reported as no change, because that is what it is. Your pages first ## Your site is the part that decides it An assistant does not read one page and answer. In one market we measured, a single answer rested on eleven pages on average — and your own pages are the ones that state the facts everything else repeats: what you do, what it costs, who you serve, how fast. That is why most of an engagement is spent on your site: service pages that answer the buying question in the first two sentences, prices and areas written plainly, structure a machine can parse. Once your pages say it clearly, the listings and profiles that quote you are made to agree — and the answer stops contradicting itself. How we build the question set and read the sources → Eleven pages behind one answer. Filled squares are pages you control and can change this week; outlined ones repeat what they say. GEO, AEO, SEO ## GEO, AEO and SEO services, side by side Generative engine optimization and answer engine optimization describe the same job under two acronyms. Both are different from search engine optimization in what they win and how they are measured — and a proposal that blurs the three is usually selling a rename. | | SEO | GEO / AEO | |---|---|---| | What you win | A position in a list of links | Your name inside a written answer | | The metric | Rank for a keyword | Share of answers that name you | | Where it is decided | Your pages and links to them | Your pages, and every place that repeats them | | Stability | The same result for everyone, minute to minute | Different on every ask — which is why we repeat each question | | Where the data lives | Search Console and analytics | Only where somebody asks and records it | | Attribution | A click you can see | Often a name remembered and typed in later | | Time to move | Weeks to months | Weeks to months, from the next crawl of the changed pages | The long version, with worked examples: GEO vs SEO. Order of work ## Measure, fix, measure again Five steps, in this order. The last one is the reason the first one matters: without a second measurement nobody can say whether the work did anything. - Step 1 ### Measure where you stand Your question set is written and run across the assistants you choose. You get the share, the names that came back instead, and every raw answer. - Step 2 ### Find why you are absent Per question type, not in general. Most businesses are visible on general recommendation questions and invisible on price and urgency ones. - Step 3 ### Publish the pages that were missing Prices, service names, areas covered — answered in the first two sentences under a heading phrased the way buyers ask. - Step 4 ### Fix the sources those answers cite The citation list is the to-do list. Listings corrected and completed, contradictions between them removed. - Step 5 ### Measure again and report what moved Eight weeks after the work, because corrected sources take time to be re-read. Reported per assistant, never averaged. 63 businesses named, 46 of them by one assistant only. An average across the four would describe none of them, so we never report one. Deliverables ## What you have on paper afterwards - Your share per assistant, with the names that came back instead - Gaps by question type — general, price, urgency, trust, comparison - Every raw answer, kept verbatim and dated - The sources behind them, listed per question - An ordered list of changes, most useful first - A log of what was published, with dates and approvals - Dashboard access from day one, with the history - A second measurement that says what moved and what did not Cost ## What AI search optimization services cost here Two things move the figure: how many assistants you want tracked, and whether we do the work or hand you the list. Tell us your market and the quote comes back the same day, by email. ### Express check Free one-off Enough to see whether you have a problem. - One assistant of your choice - Your share, and the names that came back instead - Enough to know whether it is worth going further ### Full scan from $150 one-off The whole picture, and what we would change first. - The assistants you choose, measured in depth - Gaps by question type, and by assistant - Every raw answer and the pages behind it - Dashboard access from day one - A written list of what to fix, in order ### Foundation from $300 one-off The scan, the work, and a second measurement to check it. - Everything in Full scan - The changes carried out for you, published and logged with dates - Nothing goes live without your approval - A second measurement eight weeks later to check it moved ### Monitoring from $250 per month For businesses that want the number to keep moving. - A measurement every two weeks - Dashboard with the full history - Ongoing work every month, not just reporting - Nothing goes live without your approval - An alert when your share moves enough to matter Not sure which one? Start with the free check — it costs nothing and the answer usually arrives within thirty minutes. Full prices and what moves them are on pricing. What nobody can sell you Placement in an answer. The lists are not published, not stable and not for sale. Any agency promising a position in ChatGPT is promising something it cannot control. What is sellable is the measurement, the work underneath it, and an honest account of what changed. Who this is for ## Businesses people ask an assistant to choose between Where the buyer asks a question rather than browses a list — services, clinics, professional firms, anything local with a shortlist at the end of it. - Local businesses - Clinics and healthcare - Law firms - One-off audit - Home services - B2B services - Marketing agencies, white-labelled Common questions ## Before you brief anyone How much do AI search optimization services cost? + AI search optimization services start at $150 for a one-off audit, $300 for foundation work — the audit plus the changes carried out and a second measurement — and $250 a month for continuous monitoring. The figure depends on how many assistants you want tracked, how many questions your market needs and how big your site is. Tell us your market and the quote comes back the same day by email, with no call to sit through. What is the difference between GEO, AEO and AI search optimization? + Generative engine optimization (GEO), answer engine optimization (AEO) and AI search optimization are three names for the same job: getting a business named and cited inside a generated answer rather than ranked in a list of links. We use the plainer phrase because the work is identical whichever acronym a proposal uses. Are you an agency or a tool? + Sayso is a service rather than a tool. You are not buying a dashboard and a seat — you get the measurement and the work that follows it, with the dashboard included so you can check every figure against the answer behind it. Tools report that you are absent; nobody is paid to make you present. Which assistants do you optimise for? + AI search optimization here covers four assistants — ChatGPT, Gemini, Claude and Perplexity — plus Google's AI Overviews where they appear for your questions. You choose which of them to track: ChatGPT and Claude cost several times more to query than Gemini and Perplexity, and that is the main thing that moves the price. How long before the answers change? + Assistant answers take weeks to months to change after the work. Pages we publish on your own site are live the day you approve them and are read on the next crawl; corrections to places that repeat you take longer, because somebody there has to accept them. We re-measure eight weeks after the work for that reason. Can you guarantee ChatGPT will name us? + No AI search optimization service can guarantee that ChatGPT will name you. The lists are not published, not stable and not for sale. What we can promise is that you will know where you stand, what the answers were built from, and exactly what was changed and when. Do you work with marketing agencies? + Yes — AI search optimization services here are available to marketing agencies, including white-labelled. Several of our measurements are delivered to an agency's client under the agency's name. The method and the raw answers come with it, so the report survives being questioned. What do we have to do ourselves in an AI search optimization engagement? + Your side of an AI search optimization engagement is two things: approving what gets published, and giving us access where a listing belongs to you. Everything else is ours. Clients who prefer to carry out the changes in-house take the audit on its own and work from the ordered list. ## Start with the free check Tell us your site and your city. We write the questions your customers ask, run them, and email you what the assistants said — usually within thirty minutes, with no call attached. --- ## AI Visibility Audit - Are Assistants Naming You? URL: https://ai-sayso.com/ai-visibility-audit One-off, from $150 # AI visibility audit Find out whether ChatGPT, Gemini, Claude and Perplexity name your business when buyers ask — and why they don't. One measurement, a written report, an ordered list of what to change. Yours to keep, whoever carries it out. - Back in one to two working days - Nothing to install - No retainer attached ## Share of answers, by question type ChatGPT · 5 repeats - General 38% - Service 45% - Trust 20% - Price 0% - Urgency 0% The shape of a real finding: a business in the search top three, named in a third of general answers and in none of the questions where money is mentioned. A single blended score would have read 26% and told nobody anything. Six answers ## What an AI search visibility audit tells you Each of these is a separate finding, and each one changes a different decision. A report that gives you one number instead gives you none of them. ### Are you named at all, and how often Your share of answers on each assistant, from the same questions asked several times. Not a score out of a hundred — a count you can check against the answers behind it. ### Which assistant is the problem Rarely all of them. In one market we measured, a business ran at 100% of answers on one assistant and 50% on another, on identical questions. ### Which questions you lose Visible on general recommendation questions and absent on price and urgency ones is the common shape. It is also the half of the funnel nearest a decision. ### Who is named instead of you The businesses the assistants currently trust in your category, listed in the order they came back. Usually at least one name the client had not been watching. ### What those answers were built from Every source cited, per question. That list is the reason the audit is actionable rather than interesting: it is where the work has to happen. ### What to fix first, in order A written list, most useful first, with what each item is expected to move and what it cannot move. Yours to keep, whoever carries it out. Two sizes ## The free check, and the full audit The free one is genuinely free and genuinely smaller. It answers whether you have a problem; it cannot tell you where the problem is, because one assistant and a short question set cannot show that. | | Free check free · one assistant | Full audit from $150 · what most clients buy | |---|---|---| | Assistants | One, your choice | Up to four, plus AI Overviews where they appear | | Questions | A short set, general intent | Full set across six intents, including price and urgency | | Repeats per question | Enough for a sense of it | Several, so the figure is a share and not a snapshot | | Raw answers | In the email | All of them, dated, in the dashboard | | Sources behind each answer | Not included | Listed per question | | Who is named instead | The names | The names, their share, and what cites them | | Ordered list of fixes | Not included | Written, most useful first | | Turnaround | Usually within thirty minutes | One to two working days | Both are one-off and neither puts you on a retainer. The free check tells you whether you have a problem; the full audit tells you which assistant, which questions and which sources — and what to do about it. Order of work ## How the audit runs Five steps, usually finished in one to two working days. Nothing to install and no access needed: we read your public site the way an assistant does. - Step 1 ### You tell us the site and the market The domain, the city or region, and the services that matter commercially. Nothing to install, no access to anything. - Step 2 ### We write the question set From your own pages and your market, grouped by intent: general recommendation, specific service, price, urgency, trust, comparison. - Step 3 ### We ask, and we ask again Each question several times on each assistant you chose. Identical asks return different answers, so a single run is an anecdote. - Step 4 ### We read the answers and the citations Who was named, in what order, and which pages the answer rested on. Every answer is kept verbatim and dated. - Step 5 ### You get the report and the ordered list Yours to keep and to hand to whoever does the work, including an in-house team or another agency. By question type ## Most businesses are visible and invisible at the same time Audited as one number, a business looks averagely present. Audited by what the buyer was asking, the same business is usually fine where nothing is at stake and absent exactly where money is mentioned. That pattern is common enough to expect: the questions nearest a decision — what does it cost, who can see me today — are answered from sources that list prices and opening hours, and most businesses have never checked what those sources say about them. See a real report, open in full → One business, one assistant, five intents. The two zeroes are the audit's actual finding; the 26% average would have been the only number in most reports. What it is not ## What the audit is not ### Not a ranking There is no position inside a generated answer. Any audit reporting "position 4" for an assistant is reporting an invention, and a report built on an invented number is worse than none. ### Not a promise of placement The lists are not published, not stable and not for sale. What the audit buys you is knowing where you stand and what the answers were built from. ### Not only a crawl of your site Technical and content checks on your pages are the core of it, and they are read against the answers themselves — plus every source those answers cited. A crawl alone cannot tell you which of your pages a buying question actually needed. ### Not a subscription One measurement, one report, no retainer. If you want the number watched over time that is monitoring, and it is a separate decision you make afterwards. Worth knowing before you buy any audit Ask to see the raw answers. A share that cannot be traced back to the text that produced it is not checkable, and the first time a colleague opens ChatGPT and sees something different, the whole report goes in the bin. Ours ships with every answer kept verbatim and dated, and the method is published in full so it can be copied by anyone, including your own team. Common questions ## Asked before most audits are bought What is an AI visibility audit? + An AI visibility audit is a measurement of whether AI assistants name your business when people ask the questions your customers actually ask, plus a reading of why. It covers your share of answers on each assistant, the businesses named instead of you, the sources those answers were built from, and an ordered list of what to change. It is a one-off exercise with a report at the end, not a subscription to a dashboard. How much does an AI visibility audit cost? + An AI visibility audit starts at $150. The figure depends on how many assistants you want covered, how many questions your market needs and how big your site is. ChatGPT and Claude cost several times more to query than Gemini and Perplexity, which is the main thing that moves it. Tell us your market and the quote comes back the same day by email. Is there a free AI visibility audit? + The free version of an AI visibility audit is our free check, and it is honest about being smaller: one assistant of your choice, a short question set, your share and the names that came back instead. It is enough to tell whether you have a problem worth paying to understand. It is not enough to tell you which assistant or which question type is costing you, because one assistant cannot show you that. How long does the audit take? + An AI visibility audit usually takes one to two working days from the moment you tell us the market. The free check comes back the same day, usually within thirty minutes. Which assistants does the audit cover? + An AI visibility audit covers ChatGPT, Gemini, Claude and Perplexity, plus Google's AI Overviews where they appear for your questions. You choose which to include. Each is reported separately and never averaged: a single blended score would hide the assistant you are absent from, which is the one worth knowing about. Can you audit what assistants say about our brand, not just whether we are listed? + Yes — an AI visibility audit records how your business is described when it is named, not only whether it is named. Alongside the share, the report records when it is named — the services attributed to you, the prices quoted, the claims made on your behalf, and anything factually wrong. Wrong facts in an answer usually trace back to a source that is out of date, and the citation list says which one. Do we need a Wikipedia page to be visible in AI search? + A Wikipedia page is not needed to be visible in AI search, and chasing one is a poor use of the budget for most businesses. What matters is being described accurately and consistently on the sources an assistant actually cites for your questions, which are usually directories, review platforms and local roundups. The audit tells you which ones those are for your market rather than assuming. Who should run the audit — can our agency do it? + Anyone can run an AI visibility audit — your agency included — if they ask the questions, repeat them, record the raw answers and report per assistant. Ask your supplier for those four things. If what comes back is a single score with no answers behind it, it was not measured — and the method we use is published in full so it can be copied. What happens after the audit? + What happens after an AI visibility audit is your choice. The list is yours, and plenty of clients carry it out in-house or hand it to an existing agency. If you would rather we did it, foundation work starts at $300 and includes a second measurement eight weeks later to check whether anything moved. ## See the small version first The free check runs one assistant on your market and emails you the answers, usually within thirty minutes. If the result is boring, you have saved $150 and twenty minutes. If it is not, you will know exactly what the full audit is for. --- ## AI Search Optimization for Local Business URL: https://ai-sayso.com/local-business For businesses with a catchment area # AI search optimization for local business When somebody asks an assistant who to use nearby, three or four businesses get named and the rest of the street does not exist. We measure whether you are one of them, fix the local sources those answers are built from, and measure again. - Any English-speaking market - No call, no retainer - Free result in about thirty minutes Assistant answer — example Who's good near me and can take someone on this week? Based on what's listed locally, Harbour Street is taking new customers, Northgate lists evening slots, and Mill Lane publishes its prices. I couldn't confirm availability for the others. - business profile - local directory - review platform - city roundup "I couldn't confirm" is what absence looks like: the fact was not on your pages, so nothing could repeat it. What is different locally ## Four things that do not work the way local SEO taught you ### Nobody asks a local question the way they type one Buyers type “dentist odesa” into a search box and ask an assistant “who can see me today without an appointment?”. The phrasing carries the urgency, and the urgency decides which businesses get named. ### The practical facts have to be published somewhere Opening hours, prices, areas covered, whether you take new customers. These belong on your own pages first, where you control them, and then on the listings that repeat them. A fact that exists in neither place cannot be quoted by anybody. ### One assistant is not the market Four assistants asked the same local questions named 63 businesses between them, and only 4 appeared on all four. Checking the one you happen to use tells you about a quarter of your market. ### A share, not a ranking There is no position inside a written answer. What there is: how often you are named out of how many asks, per assistant, per question type. Small enough to count by hand, which is why you can check ours. Six source types ## Where a local answer comes from Six source types do the work in a local market, and your own site sets the facts for all of them. This is the list we work down after a measurement, starting with your pages. | Source | What it answers for the assistant | What usually goes wrong | |---|---|---| | Your own site | Prices, specifics, service names, areas covered — the detail nobody else publishes | The answer buried in paragraph four, or no prices at all | | Your business profile | Hours, location, category, whether you are open now | Categories left at the default, hours out of date after a change | | Local directories | Services offered, areas covered, contact details | A listing created years ago, with half the services missing | | Review platforms | Whether you are any good, and at what specifically | Reviews that never mention the service you want to be asked about | | City roundups and "best of" posts | The shortlist itself — these are often quoted almost verbatim | You are not on them, and nobody has ever asked to be | | Forums and community threads | What locals actually say when somebody asks for a recommendation | Nothing to do about it directly, and worth knowing it exists | In one market we measured, a single answer rested on eleven pages on average. Your own site is the one that decides what the other ten can say. Three businesses across six local sources. Filled means the source describes them; empty means it does not. The empty squares are the whole job, and they are cheaper to fix than anything on your website. Why it is fixable ## Most of it is on pages you control Local businesses are usually absent for an unglamorous reason: the facts a buyer asks about — price, availability, area covered — are not written anywhere a machine can lift them, starting with their own site. That is good news commercially. Publishing a clear answer on your own page is work we can do this week, and making the four or five sources that repeat you agree with it is slow but certain. Unlike a ranking, neither has to be won from a competitor. Order of work ## From first measurement to second - Step 1 ### Your questions, in your city Written from your services and your area, including the ones that mention price, opening hours and “today”. Those are the questions local buyers actually ask and the ones most businesses lose. - Step 2 ### Measured on the assistants you choose Each question several times on each assistant, with every answer kept. You see your share, the names that came back instead, and what each answer cited. - Step 3 ### The pages a local buyer needs, published Prices and areas covered, answered in the first two sentences under a heading phrased the way people ask. Nothing goes live without your approval. - Step 4 ### The listings those answers rely on, corrected Same services, same hours, same prices everywhere. Contradictions between sources are a reason for an assistant to name somebody else. - Step 5 ### Measured again Eight weeks later, same questions, same method. Movement under about five points is reported as no change. Who we work with ## Local businesses people ask an assistant to choose between Anything with a shortlist at the end of the question. Two sectors have pages of their own because what patients and clients ask is specific enough to need them. - Clinics and healthcare - Law firms - Dental practices - Veterinary clinics - Accountants - Home services and trades - Salons and studios - Estate and letting agents - Driving and training schools - Restaurants and venues Markets We work remotely and measure a city by asking its questions, so the market is wherever your customers are — New York, Los Angeles, Houston, Austin, San Antonio, San Diego, Long Island and anywhere else English-speaking. The assistants answering those questions have never been to your city either. Common questions ## What local owners ask first What is the role of AI in local SEO now? + AI has added a second surface to local SEO rather than replacing the first. People still search and still click maps results; a growing share also ask an assistant to shortlist for them, and that answer is assembled from directories, review platforms and local roundups rather than from a ranking. The practical consequence is that local SEO work — consistent listings, real reviews, accurate hours — now pays twice, and that there is a second number to measure which no rank tracker reports. Does this replace my Google Business Profile work? + AI search optimization for a local business does not replace Google Business Profile work, it depends on it. A complete, accurate profile is one of the sources these answers are built from. What it does not do is tell you whether ChatGPT or Perplexity names you, because neither reads your profile the way Google's map pack does. We only have one location. Is this worth it? + AI search optimization is often worth more to a single-location business than to a chain. Single-location businesses are the ones assistants struggle to place, because there is less written about them elsewhere — and that is also why a small amount of work on the right four or five sources moves the number. Do you work with businesses outside your own city? + Yes — we work with local businesses in any English-speaking market. The work is remote and the market is wherever your customers are: we measure a city by asking its questions, not by being in it. Most of our measurements are for markets we have never set foot in, which is also true of the assistants answering them. How much does AI search optimization for a local business cost? + AI search optimization for a local business starts at $150 for a one-off audit and $250 a month for continuous work. For a single-location business with one service area, the audit is usually the right first purchase: it is cheap enough to be a decision rather than a project, and it tells you whether anything further is worth buying. How do I know the answers are not just different for everyone? + Assistant answers to local questions genuinely are different every time — that is the point of asking each question several times. One ask is an anecdote; a share across repeats is a measurement. Anyone reporting a local business's position in an assistant answer is reporting something that does not exist. Can we measure AI visibility ourselves? + You can do part of the measurement yourself: open an assistant, ask your five most valuable buying questions, write down who was named. Twenty minutes, and it will tell you more than any article. What is hard to do by hand is the repetition, keeping the answers, and staying honest — everyone remembers the answer that named them and forgets the three that did not. ## Ask your city's question for free Tell us your site and your area. We write the questions your customers ask, run them, and email you who got named — usually within thirty minutes. If it is your competitors, at least you will know by lunchtime. --- ## AI Visibility by Industry - Who We Work With URL: https://ai-sayso.com/industries Industries # AI visibility, by what your buyers ask The method is the same everywhere: ask the questions buyers ask, repeat them, read the sources behind the answers. What changes by sector is which questions matter commercially and which sources decide them — and those two things decide where the work goes. ## Clinics and healthcare Patients ask symptom-first and price-second, and assistants are visibly careful with health questions. Which sources they lean on instead is the whole game. See the page → ## Law firms Nobody asks for a solicitor in general — they ask about a situation. The firms named are the ones described by practice area on sources outside their own site. See the page → What a sector changes ## Three things change, one does not - Which questions carry money. For a clinic it is availability and price; for a law firm it is the matter type; for a restaurant it is the occasion. The question set is written per sector, and it decides everything downstream. - Which sources answer them. Legal directories matter in one sector and do not exist in another. The citation list from the first measurement tells you, rather than a template telling you. - How cautious the assistants are. Health, legal and financial topics get hedged answers that name nobody. That ceiling is real, it is measurable, and it is better known before a budget than after one. - What does not change: the method. Same questions repeated, same answers kept verbatim, same reporting per assistant, published in full on how it works. Not listed Most of our work is for businesses with no sector page here — trades, agencies, B2B services, anything with a shortlist at the end of the question. The general version is local business, and the service itself is described on AI search optimization services. If you want to know whether your category is one assistants answer at all, the free check answers that in about thirty minutes. ## Find out what your market is told Whatever the sector, the first step is the same: ask the questions your buyers ask and see who gets named. The free check does that for one business and emails you the answers, usually within thirty minutes. --- ## AI Search Optimization for Healthcare Clinics URL: https://ai-sayso.com/industries/healthcare # AI search optimization for healthcare Patients ask assistants who can see them, what it costs and whether anyone takes new patients. We measure whether your clinic is named in those answers, fix the listings and pages they are built from, and measure again. - No patient data involved - Nothing clinical written by us - Dental, vet and private practice ## Questions a clinic is measured on Excerpt - My filling fell out — who can see me today? - Which clinics near me are taking new patients? - How much does an implant cost privately? - Who is good with anxious patients? - Is there a clinic open on Saturday around here? Six intents, asked several times each, on every assistant you choose. The clinic that answers these questions publicly is the clinic that gets named. What decides it here ## Healthcare answers behave unlike any other category Everything below changes what is worth doing — and, just as usefully, what is not. ### Patients describe a problem, not a service “My filling fell out, who can see me today?” is the question. It names no service and no clinic, and the assistant has to decide what it is about before it decides whom to recommend. ### Assistants are careful with health On medical questions they hedge, decline to recommend anyone, or stay with sources they consider safe. That caution is itself measurable, and it changes what is worth working on. ### The practical facts decide the answer Opening hours, whether you take new patients, whether you see children, what an appointment costs. Few clinics publish all of it on their own pages, and fewer still make the listings repeat it without contradiction. ### Nothing clinical is written by us We work on listings, service descriptions, prices and structure. Anything that makes a clinical claim is written or approved by your clinicians, not by a marketing supplier. Patient intent ## What a patient asks, and what answers it Every one of these is answered from a fact somebody published. Your own pages are where that fact should start, and the listings should repeat it word for word. | What the patient asks | What the assistant needs | Where it is read from | |---|---|---| | Can somebody see me today? | Opening hours, emergency slots, whether you take walk-ins | Your own site, then business profile and directories | | Are you taking new patients? | A stated yes or no, dated | Your own site, repeated in directory listings | | What does it cost? | A figure or a range it can quote | Your price page, directories, review threads | | Do you treat this specific thing? | Service names in the words patients use | Service pages, directory categories | | Are they any good? | Reviews that mention the treatment, not just the parking | Review platforms, against what your pages claim | | Who is the clinician? | A named person with visible qualifications | Your team page, professional registers | Where the ceiling is ## Some health questions name nobody, and that is the answer Assistants treat medical topics carefully. Ask about symptoms and most of them decline to recommend a practice at all, suggesting a professional in general terms. Ask about an appointment, a price or availability, and the same assistant names three clinics without hesitating. A measurement that reports both tells you where your budget can do something and where it cannot. Nobody can sell you visibility on a question the assistant has decided not to answer — and a supplier who does not mention this has not measured it. Why regulated topics behave differently → The shape healthcare measurements tend to take: practical questions name clinics freely, symptom questions barely name anyone at all. Order of work ## From first measurement to second - Step 1 ### Patient questions, written from your services Grouped the way patients think: symptom, treatment, price, urgency, suitability, trust. Not keywords — the sentences people actually type into an assistant at eleven at night. - Step 2 ### Measured across the assistants you choose Including how often the assistant declines to name anyone at all, which in healthcare is a result in itself and tells you where the ceiling is. - Step 3 ### The pages patients are sent to, rewritten The answer in the first two sentences under the heading, with your clinicians named and their qualifications where a reader can see them. - Step 4 ### Listings and profiles made to agree Hours, services, whether you are accepting patients, prices where you publish them. One set of facts, the same on every source an answer cited. - Step 5 ### Measured again, eight weeks later Same questions, same assistants. Movement reported per assistant, and anything under about five points reported as no change. Where we stop We do not write clinical content, and we do not touch patient data. Treatment descriptions, outcomes and anything a reader could act on medically are written or signed off by your clinicians, with their name and qualifications on the page. Everything we work on is public: listings, service names, prices, availability, structure. Where it works best ## Where this works best Practices whose patients ask practical, commercial questions — cost, availability, suitability — rather than clinical ones. - Dental practices - Veterinary clinics - Private GP practices - Physiotherapy and osteopathy - Dermatology and aesthetics - Optometry - Fertility and women's health - Mental health practices - Any local practice Common questions ## What practice managers ask Do AI assistants even recommend clinics? + AI assistants do recommend clinics, and more cautiously than they recommend restaurants. For practical questions — who takes new patients, who is open on a Saturday, what a treatment costs — they name specific practices readily. For anything closer to diagnosis they hedge, suggest seeing a professional, and name nobody. Both outcomes are measurable, and the second one tells you which questions are not worth chasing. We are a private clinic. Which questions matter commercially? + For a private clinic the commercially important questions are price and availability. In every measurement we have run, those are where businesses are most often absent and where the buyer is closest to booking. Clinics that publish prices and new-patient availability clearly — on their own site and on the directories an answer cites — appear in answers that clinics with better websites do not. Will you write medical content for us? + We do not write medical content. AI search optimization for healthcare as we do it stops at the non-clinical layer: service names, prices, availability, structure, listings. Anything that makes a clinical claim is yours to write or approve, with the clinician named. That is both the right thing and the thing retrieval systems reward — attributed, qualified authorship is part of how a health source is weighed. Does this apply to veterinary and dental practices too? + AI search optimization for healthcare applies to veterinary and dental practices too, and both tend to show stronger results than general medical practices for the same reason: owners and patients ask practical, commercial questions — cost of a procedure, emergency availability, whether a breed or a treatment is handled — and assistants answer those without the caution they apply to symptoms. Is any of this a problem for patient privacy or compliance? + AI visibility work for a clinic touches no patient data and raises no compliance question of its own. We work entirely on public sources: your public pages, public directories, public review platforms. The measurement itself is just questions asked of an assistant, with the answers kept. How much does it cost for a clinic? + AI search optimization for a clinic starts at $150 for the audit, which tells you where you stand across the assistants you choose. Foundation work — the audit plus the corrections and a second measurement — starts at $300. Multi-site groups cost more because each location has its own listings and its own answer. Can you guarantee an assistant will recommend our clinic? + No supplier can guarantee that an assistant will recommend your clinic, and in healthcare be especially wary of anyone who does. Assistants deliberately decline to recommend on some health questions, and no supplier controls that. What is controllable is being described accurately and completely wherever these answers are read from. ## Find out what patients are told Tell us the clinic and the area. We ask the questions your patients ask and email you the answers, usually within thirty minutes. No call, no obligation, and the raw answers come with it. --- ## AI Search Optimization for Law Firms URL: https://ai-sayso.com/industries/law-firms # AI search optimization for law firms Somebody describes their situation to an assistant and asks who handles it. Two or three firms get named. We measure whether yours is one of them, correct the directories and listings those answers rest on, and measure again eight weeks later. - Measured per practice area - Nothing published without approval - No outcome claims, ever Assistant answer — example I was dismissed after raising a complaint. Who deals with this locally? That sounds like an employment matter. Calder & Rowe lists unfair dismissal work, Hartley Legal offers a fixed-fee first consultation, and Bellbrook is listed for employment tribunals. - legal directory - review platform - regional roundup - firm website Three firms named out of a city full of them. The deciding words — "unfair dismissal", "fixed fee", "tribunal" — have to be on your own pages before anything can repeat them. What decides it ## Four things that decide whether a firm gets named ### Nobody asks for a lawyer in general They describe a situation — a dismissal, a boundary dispute, a visa refusal — and ask who handles it. The firms named are the ones whose practice areas are described in words that match the situation. ### Legal is a cautious category Assistants hedge on legal questions and Google shows AI Overviews for them inconsistently. Where they do name firms, they lean on directories and professional listings more heavily than in other sectors. ### Practice-area pages do the deciding A page per matter type, in the words clients use, is what makes a firm a candidate at all. Professional directories, review platforms and association listings are read next, and they should repeat exactly what those pages say. ### Fee questions are where firms disappear Fixed fees, hourly rates, no-win-no-fee terms. The firms named on cost questions are almost always the ones that publish something, anything, about price. How instructions start ## The situation, not the specialism Clients arrive with a problem and no vocabulary for it. The assistant translates — and whether it translates towards your practice areas depends on how they are described where it reads. | What the client says | What it is | What gets a firm named | |---|---|---| | "They let me go after I complained" | Unfair dismissal, employment | The phrase "unfair dismissal" in your listed services, not just "employment law" | | "My ex won't agree about the children" | Children arrangements, family | Practice area listed by matter, plus visible fee structure | | "My visa was refused" | Immigration appeal | Jurisdiction, languages spoken, appeal work stated explicitly | | "The sale is falling through" | Conveyancing | Fixed fees published somewhere an answer can quote | | "I was hurt at work" | Personal injury | No-win-no-fee terms described plainly, reviews mentioning the matter | | "My father died without a will" | Probate | Service named the way people say it, not "contentious estates" | We write the question set from the matters you actually take on, per office and per jurisdiction. Three firms across six sources an answer cited, starting with their own sites. Empty squares are where the matter type is not described at all. Your pages, and what repeats them ## Your pages set the facts the rest repeat A firm whose site describes "employment law" in general loses to a firm whose site names unfair dismissal, settlement agreements and tribunal representation — and whose directory entries say exactly the same words. So an engagement starts on your own pages: one per matter type, in the client's vocabulary, with fee structure and jurisdiction stated plainly. Then every listing that quotes you is brought into line with them. Order of work ## Measured, corrected, measured again - Step 1 ### Questions by matter, not by keyword Written from the situations your clients arrive with, per practice area and per jurisdiction you serve. - Step 2 ### Measured across assistants Including how often the assistant refuses to recommend anyone and answers with general guidance instead — common in legal, and worth knowing before you budget. - Step 3 ### Practice-area pages made answerable One page per matter type, with the first two sentences answering the question a client would ask, and the named solicitor visible on it. - Step 4 ### Directory and listing entries corrected Practice areas, jurisdictions, languages, fee structures and accreditations, made consistent across every source an answer cited. - Step 5 ### Measured again Eight weeks on, same questions, reported per assistant and per practice area so you can see which of them moved. Compliance Nothing goes live without your approval, and we write no claims about outcomes. No success rates, no comparative superlatives, no guarantees — these create a regulatory problem for you and buy nothing measurable. What we publish is factual: practice areas, jurisdictions, fee structures, accreditations, named solicitors. Practice areas ## Where assistants name firms most readily Consumer-facing matters, where the client knows the problem but not the specialism. Measured separately — a firm can be strong in one and absent in the next. - Employment - Family and children - Immigration - Personal injury - Conveyancing - Wills and probate - Landlord and tenant - Criminal defence - Local firms generally Common questions ## What partners ask before buying this Do people really choose a law firm through ChatGPT? + People shortlist law firms through ChatGPT rather than instruct through it. The pattern we see is an assistant used to understand the situation and narrow to two or three firms, then ordinary search used to check them. That makes the assistant answer the step that decides who gets considered at all — and the one nobody is measuring. Which practice areas benefit most? + AI search optimization for law firms works best in practice areas where the client knows the problem but not the specialism: employment, family, immigration, personal injury, conveyancing, wills and probate. Corporate and commercial work moves less, because those instructions come through referral rather than through a question typed at midnight. What about advertising rules and claims? + Everything we publish for a law firm stays inside your regulator's advertising rules, and nothing goes live without your approval. We do not write comparative claims, success-rate statistics or anything resembling a guarantee of outcome — in this sector those create a compliance problem and no measurable benefit. Do we have to publish our fees? + A law firm does not have to publish fees, and firms that publish nothing about cost are named less often on fee questions than firms that do. A fee range, a fixed-fee list, or an honest explanation of what drives the cost is usually enough — an assistant answering a price question needs something to quote, and silence is the reason it quotes somebody else. Our rankings are strong. Why would assistants ignore us? + A law firm can rank well and still be ignored by assistants, because a ranking is earned by your pages and an answer is assembled from everybody else's. We measured a business sitting in the search top three that appeared in 38% of general answers and 0% of price answers — strong where its own site decided the outcome, absent where other sources did. How much does it cost for a law firm? + AI search optimization for a law firm starts at $150 for the audit, which covers the practice areas you choose. Foundation work starts at $300 and includes the corrections and a second measurement. Firms with several offices or jurisdictions cost more, because each one is a separate market with separate answers. Can you guarantee our firm will be recommended? + No one can guarantee that an assistant will recommend your firm. The lists are not published, not stable and not for sale, and in a regulated category assistants decline to recommend anyone more often than elsewhere. What we can do is make sure that when a firm is named, there is no good reason for it not to be yours. ## See which firms your clients are told about Give us the firm, the area and one practice area that matters. We ask the questions clients ask and email you the answers, usually within thirty minutes — including the firms named instead of yours. --- ## AI Assistant Recommendations: Local Market Study URL: https://ai-sayso.com/research/ai-assistants-local-recommendations-2026 Original research · dental practices in Odesa, Ukraine # What four AI assistants answer when a local customer asks for a recommendation Two measurements on one local market, 239 answers in all. Who gets named, how little the assistants agree, what the answers are built from, and where visibility collapses. Every figure below was recomputed from the stored answers on 1 October 2026 and can be checked against them. By Olha Kryva Measured 7 September 2026 and 8 September 2026 - 63 different businesses named across 24 answers to the same three questions - 4 of them named by all four assistants. The rest depend on which one your customer opens - 11.2 pages cited behind an average answer, from 57 different domains - 0% share for one practice on urgent questions, against 92% on general ones Method ## Two measurements, one market One goes wide across assistants, the other deep into one. Both were run on dental practices in Odesa, Ukraine, with questions written the way customers ask them, in Russian and Ukrainian. | | Breadth | Depth | |---|---|---| | Date | 8 September 2026 | 7 September 2026 | | Assistants | ChatGPT, Gemini, Claude, Perplexity | Gemini | | Questions | 3 | 43 | | Repeats per question | 2 | 5 | | Answers recorded | 24 | 215 | | Business mentions extracted | 183 | 1836 | Every answer was stored verbatim with its citations. Nothing was asked once: a single answer is a coin toss, and the figures below are shares across repeats. Finding 1 ## Four assistants, almost no agreement The same three questions, asked twice on each of four assistants, produced 63 different business names in 24 answers. 4 of those businesses were named by all four assistants. 46 were named by exactly one. A single blended "AI visibility score" would average four lists that barely overlap. For a business in this market, being visible on Gemini says close to nothing about being visible on Claude — which is why every figure we report is per assistant. 63 businesses named in total; the small circle is the 4 named by all four assistants. Finding 2 ## An answer is assembled from about eleven pages 268 citations across 24 answers — 11.2 per answer, from 57 domains. The clinics' own sites account for 47% of them; directories, rating sites, booking platforms and local media for the other 53%. | Domain | Citations | What it is | |---|---|---| | top20.ua | 33 | City directory | | triodent.od.ua | 30 | A clinic's own site | | doc.ua | 18 | Medical directory | | likarni.com | 15 | Medical directory | | phildent.od.ua | 13 | A clinic's own site | | kichukdentalclinic.com.ua | 12 | A clinic's own site | | uba.top | 8 | Rating site | | stomatolohii.com | 8 | Dental directory | | odesa-day.com | 6 | Local media | | dymovdental.com | 6 | A clinic's own site | The six most-cited domains supplied 45% of all citations. The single most-cited source was a city directory; second was a clinic's own website. What this means in practice Your own pages are not a minority player here — they are the largest single category of source, and the one you write yourself. The practical order follows from that: state the facts a buyer asks about on your own pages first, then make the directories and rating sites that get read alongside them agree. A contradiction between four descriptions of the same clinic is a reason for an assistant to name a fifth. One practice, 215 answers on Gemini. Share of answers naming it, by what the customer was asking. Finding 3 ## Visibility collapses as the question gets closer to a booking Taking one practice from the set and reading its share by question type gives the shape we now expect to find: strong where nothing is at stake, absent where money or urgency is mentioned. Averaged into one number this practice sits at 57% and looks healthy. The number that matters commercially — can somebody who needs a dentist today find it — is zero. By question type ## The same practice, six kinds of question | What the customer asked | Times asked | Times named | Share | |---|---|---|---| | General recommendation | 25 | 23 | 92% | | A specific service | 75 | 53 | 71% | | Reviews and trust | 35 | 19 | 54% | | Price | 25 | 7 | 28% | | Near me | 25 | 5 | 20% | | Needed right now | 20 | 0 | 0% | Limitations ## What this study does not show It is one market, two days and 239 answers. Treat it as a worked example of a method, not as a law of nature. ### One city, one niche Dental practices in a single city. Other categories — regulated, national, B2B — behave differently, and we say so rather than generalising from one market. ### Counting names is a judgement 63 is the count when a name we could not match to a known business is counted as itself. Merge strictly and the same answers give 33. We publish both, because the method changes the headline by a factor of two. ### Three questions across assistants The breadth measurement is deliberately small: the same three questions, twice each. It is enough to show disagreement, not enough to rank assistants against each other. ### Answers move These are two days in September 2026. Assistants change their retrieval and their sources, which is the reason measurement is repeated rather than done once. A correction worth making in public This site first published 65 businesses, three named by all four and 48 named by one. Those were the figures from the first pass over the same run. Duplicate names have since been merged, and recomputing on 1 October 2026 gives 63, 4 and 46. Every page has been updated to the new figures, and this note stays here rather than the change happening quietly — a number that moves without explanation is exactly what we tell clients not to accept. ## Run this on your own market The method is published in full and copyable. If you would rather we did it, the free check runs a short version on one assistant and emails you the answers, usually within thirty minutes. --- ## Is Sayso Accurate? Method, Error and Limits URL: https://ai-sayso.com/is-sayso-accurate Method and limits # Is Sayso accurate? Accurate about what it measures, within limits we publish instead of hiding. This page sets out exactly what the number is, how it is produced, where the error comes from, how big each source of it is, and what we do about each one. Nothing here is a selling point we would drop in a bad quarter. The number ## A share of answers, per assistant Everything we report is the same simple fraction: how often an assistant named your business out of the number of times we asked. ### What it is Times named ÷ times asked, calculated separately for each assistant and each type of question. A count you could do by hand, which is the point: you can check it. ### What it is not Not a rank — there is no position inside a written answer. Not a score out of a hundred. Not an average across assistants: four numbers that disagree do not become one number that is right. ### Why it is checkable Every answer is stored word for word, dated, with the sources it cited. Any figure in a report opens into the text behind it. A share you cannot trace is a claim, not a measurement. Sampling error ## What a share from five asks can honestly mean This is the uncomfortable table, and we would rather you saw it from us than worked it out later. These are 95% confidence intervals: the range the true share could plausibly sit in, given how many times we asked. | Repeats | Result | Reported share | Honest range | |---|---|---|---| | 5 | Named 0 times | 0% | 0–43% | | 5 | Named 1 times | 20% | 4–62% | | 5 | Named 3 times | 60% | 23–88% | | 5 | Named 5 times | 100% | 57–100% | | 20 | Named 5 times | 25% | 11–47% | | 20 | Named 10 times | 50% | 30–70% | | 20 | Named 15 times | 75% | 53–89% | | 20 | Named 20 times | 100% | 84–100% | Wilson intervals, calculated rather than estimated. The practical reading: at five repeats a share tells you roughly where you stand, not precisely; at twenty it starts to be a number you can defend in a meeting. What we do about it We never report a share from a single ask, we say how many repeats a figure rests on, and we raise the count when a decision depends on it. Where the range is wide, the useful finding is usually not the percentage at all — it is which questions named you and which did not. Everything that can go wrong ## Seven sources of error, and the answer to each A limitation with no answer beside it is just a worry. Each of these costs something; each has a rule that keeps it from quietly becoming your result. | Source of error | How big | What we do | |---|---|---| | Sampling: how many times we asked | The biggest one | Five repeats of a question cannot pin a share down tightly, and we say so in the table below rather than reporting 60% as if it were exact. We raise the repeat count where a decision depends on the figure, and we never report a share from a single ask. | | Drift: the system answers differently anyway | A few points, run to run | Assistants return different answers to identical questions, with nothing changed on anybody's site. Movement under about five points is reported as no change, even when it would look like progress. | | Name matching: the same business written four ways | Can move a count by half | Assistants write business names loosely. Merging them is a judgement, and it matters: the same run gives 63 distinct businesses counted loosely and 33 counted strictly. Unmatched names go to a review queue a person works through, and the counting rule is published with the figure. | | Personalisation: the assistant knows its user | Unknown, and real | Chat history, account and location colour what an assistant says to a given person. We measure from a clean session with no history, which is closer to a new customer than to you — and is the reason your own ask will not match our report exactly. | | Model changes under us | Occasional, sometimes large | Providers update models and retrieval without notice. We record which model answered each question, so a jump that coincides with a model change is visible as what it is rather than reported as your result. | | Refusals: when nobody is recommended | Common in health, legal, finance | On some questions assistants decline to name anyone. That is recorded as a refusal, not as a zero for you — a ceiling in the category is not a failure of your business, and no supplier can sell you past it. | | Missing citations | Varies by assistant | Some assistants show their sources and some do not. Where they do not, the source list for that answer is incomplete, and the report says so instead of implying we know what was read. | Boundaries ## What we never claim ### That we can put you in an answer The lists are not published, not stable and not for sale. Anyone promising placement in ChatGPT is promising something they do not control. ### That our work caused a movement A report says what was changed, when, and what moved afterwards — and whether the movement was concentrated in the question group the work targeted. That is evidence, honestly labelled, not proof. ### That one number describes you We do not publish a blended “AI visibility score”. Four assistants that disagree produce four numbers, and the one you are absent from is the one worth knowing about. ### That we know what an assistant will do next Models and retrieval change without notice. We record which model answered and re-measure rather than predicting. A correction we made in public ## We published a number, then found it had moved Our first pass over one measurement reported 65 businesses named across four assistants, three of them by all four. After duplicate names were merged, the same stored answers gave 63 and four. Every page now carries the new figures, and the study explains the change. We are telling you this on the page about accuracy on purpose. A supplier whose numbers never change is not more accurate — they are just not recounting. Read the study, with the method and the limits → ### Same run, recounted 1 October 2026 | | First pass | Recount | |---|---|---| | Businesses named | 65 | 63 | | Named by all four | 3 | 4 | | Named by one only | 48 | 46 | Nothing about the market changed. The counting of duplicate names did, and that is a methodological fact worth publishing. Verification ## How to check us, in four moves - Open any figure. Every share in the dashboard links to the answers behind it, kept verbatim and dated. - Ask the question yourself. Expect a different answer — then ask it five times and watch the list move. - Read the method. It is published in full on how it works, including how questions are written. - Have it repeated. Ask us to re-run a measurement, or hand the method to another agency and compare. Common questions ## Asked by people who check their suppliers Is Sayso accurate? + Sayso is accurate about what it measures, and the measurement has limits we publish rather than hide. What we report is a share: how often an assistant named your business out of the times we asked, per assistant and per question type, with every answer stored word for word so any figure can be traced back to the text that produced it. What we cannot do is report a precise number from a handful of asks, predict what an assistant will say tomorrow, or tell you what it says to a specific person with their own chat history. Why did I ask the same question and get a different answer? + Because that is how these systems work, and it is the single most common question about an AI visibility report. Assistants return different answers to identical questions, and they personalise by account, history and location. We measure from a clean session with no history, which is closer to what a new customer sees than what you see. If your own ask matched our report exactly every time, the report would be the suspicious thing. How many times do you ask each question? + We ask each question five times as a floor, and every figure says how many repeats it rests on. Five is enough for a share to mean something and not enough for it to be precise: a 3-out-of-5 result honestly means somewhere between 23% and 88%, which is why we raise the count when a decision rests on the figure and why we never report a share from a single ask. The table on this page gives the arithmetic in full. What counts as real movement between measurements? + Real movement between measurements is more than about five points on the same assistant, asked the same way. Below that we report no change, even when the direction is flattering. Reporting drift as progress is the fastest way to make the rest of a report untrustworthy. Can I check your figures myself? + You can check every figure yourself, and the dashboard is built for exactly that. Every share links to the answers behind it, kept verbatim and dated, with the sources each answer cited. The method is published in full on how it works, so your own team or another agency can repeat the measurement and compare. Do you guarantee that an assistant will name our business? + We do not guarantee that an assistant will name your business, and nobody else can either. The lists are not published, not stable and not for sale. We also do not claim that work we did caused a movement: a report says what was changed, when, and what moved afterwards, and notes whether the movement was concentrated in the question group the work targeted. What happens when an assistant refuses to recommend anyone? + When an assistant refuses to recommend anyone, the answer is recorded as a refusal rather than as a zero against your business. Health, legal and financial questions get hedged answers that name nobody, and that ceiling belongs to the category. Knowing which of your questions hit it is useful: it is budget you now know not to spend. Have you ever got a published number wrong? + Yes — we published a number that later proved wrong, and the correction is on the site. Our first pass over one measurement reported 65 businesses named, three by all four assistants. After duplicate names were merged, recomputing gave 63 and four. Every page now carries the new figures, and the research page explains the change rather than letting the number move quietly. ## See the measurement on your own market The free check runs a short version on one assistant and emails you the answers, with the raw text included. Judge the method on your own business rather than on this page. --- ## About Sayso - Who Measures AI Visibility and Why URL: https://ai-sayso.com/about About # Who we are, and what we are careful about Sayso measures whether AI assistants name your business, and does the work to get you named. That is the whole product. What we hold ourselves to - Every number is traceable Each figure links to the answer it came from. You can open the assistant and check us. - We name the error bars We ask more than once, report a share, and call a small swing no change. - Assistants are never averaged A single blended figure hides the one you are absent from. Behind it is a search specialist, not a dashboard vendor. What gets a business named in an answer is the same craft that got it found in results — the listings, the directories, the round-ups an assistant reads. What changed is where the answer appears, not how the work is done. What we can do is measure honestly, show you the raw material, and tell you when something has not worked. Every account is run by a person who knows it: when a number moves, they can tell you what moved it. ## Founder Sayso was started by Olha Kryva, an SEO and GEO specialist who has spent the last decade getting businesses found in search — five years agency-side, and five before that running content for an online retailer. She teaches the SEO module on an internet marketing course at Hillel IT School, including the part about optimising for AI search: Google AI Overviews, ChatGPT, Perplexity. That is where this started — the questions students asked had no good answers yet, and neither did the tools. Every client is handled by someone who knows their account, not by whoever picks up the ticket next. Olha Kryva on LinkedIn → - 10 years in search, agency and in-house - GEO teaches AI-search optimisation at Hillel IT School - Technical SEO audits, site architecture, structured data - Local search Google Business Profiles and map visibility - Content strategy briefs, optimisation, editorial planning - Certified Google Analytics 4, Google Ads, Google Business Profile ## The facts, plainly - What Sayso is A service that measures whether AI assistants name a business, and does the work to get it named. - Assistants covered ChatGPT, Gemini, Claude and Perplexity, reported separately. - Who it is for Businesses whose customers choose a supplier locally — clinics, professional firms, trades, services. - Where we work Any city and language the assistants answer in. The question set is built for the market, not translated into it. - What we do not do Fake reviews, filler articles, pages written for machines, or promises about positions. - Contact hello@ai-sayso.com ## Ask us something Questions about the method get a straight answer, including the parts that do not flatter us. --- ## Contact Sayso URL: https://ai-sayso.com/contact Contact # Talk to a person Email or WhatsApp, whichever you prefer. We reply within ten minutes in working hours, and the reply comes from whoever will actually do the work. hello@ai-sayso.com WhatsApp +1 814 210 2559 If you want to see numbers before you talk to anyone, start with the free check or read a sample report. --- ## Privacy Policy URL: https://ai-sayso.com/privacy Privacy # What we collect, and what we never do with it Last updated 8 September 2026. Written to be read, not to be survived. ## The short version We collect what you send us in a form and what we measure about your business in public AI answers. We use it to run your measurement and to reply to you. We do not use your data to train models, and we do not sell it or share it with anyone except the API providers listed below. ## What we collect What you give us. Your email, website, city, what your business does, and anything you type into a message field. If you become a client, the details needed for invoicing. What we measure. The questions we ask assistants about your market, the answers they return, the businesses named in them, and the pages those answers cite. This is public information: anybody can ask an assistant the same question. What the website records. Standard server logs — the page requested, the time, the browser and the network address — kept for security and kept for thirty days. No advertising trackers, no cross-site profiling. ## Who else sees it To run a measurement we send the questions to the assistants themselves: OpenAI, Google, Anthropic and Perplexity. Those questions describe a market, not you as a person, and they do not include your account details or anything from your message to us. Email we send passes through our mail provider. Our servers are rented from a hosting company. That is the whole list. We do not sell data. We do not pass it to advertising networks. We do not publish your numbers or your name as a client without asking you first. ## Training Your data is not used to train any model, by us or by anyone on our behalf. Where an assistant provider offers a setting that excludes API traffic from training, we use it. ## How long we keep it Measurements and answers stay for as long as you are a client, plus twelve months, so that a later measurement can be compared with an earlier one. Enquiries that do not become clients are deleted after twelve months. Ask us to delete anything sooner and we will, unless we are required to keep an invoice. ## Your rights You can ask for a copy of what we hold about you, ask us to correct it, or ask us to delete it. Write to hello@ai-sayso.com and we will answer within thirty days. If you are in the UK or the EU and you think we have handled your data badly, you may complain to your data protection authority. ## Changes If this page changes in a way that matters, we will say so on the page and, for clients, by email. The date at the top always shows the current version. --- ## Cookie Policy URL: https://ai-sayso.com/cookies Cookies # The one cookie we set, and why Last updated 29 September 2026. Short, because there is not much to say. ## The short version The public pages of this site set no cookies at all. You can read every page, run the free check and write to us without one. One cookie is set when you sign in, and it exists so that the next page knows it is still you. There are no advertising cookies and no third-party trackers on this site. ## What the cookie is Name. sayso_session What it holds. Your account number and the time you signed in, signed with a key only our server has. Nothing else — not your email, not your name, nothing about your business. The signature is what stops the value being edited into somebody else’s account. How long it lives. Thirty days, or until you sign out. Signing out clears it immediately. How it is protected. It is sent only over an encrypted connection, it is not readable by scripts in the page, and it is not sent when another site links to us. ## What we do not use No analytics cookies, no advertising cookies, no pixels, no cross-site profiling, no embedded widgets that set cookies of their own. If that changes — an analytics tool is the likely candidate — this page will say so before it happens, and the choice will be yours to make. ## Server logs are not cookies Our server records the page requested, the time, the browser and the network address, and keeps that for thirty days for security. That happens for every website and does not involve a cookie or identify you by name. It is described in the privacy policy along with everything else we hold. ## Turning it off Every browser can block or delete cookies for a site. Blocking ours costs you nothing on the public pages; the dashboard will simply ask you to sign in again on every page, which makes it unusable. That is the whole trade. ## Questions Write to hello@ai-sayso.com. A person replies within ten minutes during working hours. --- ## Terms of Service URL: https://ai-sayso.com/terms Terms # The terms, including what we do not promise Last updated 8 September 2026. ## What we sell A measurement of whether AI assistants name your business in answers to buying questions in your market, a report explaining it, and — unless you have chosen to do the work yourself — the work on the sources those assistants read. ## What we do not promise We do not promise that an assistant will name you, that a share will rise, or that a position will improve. The assistants do not publish how they build their lists, they change without notice, and no supplier can control them. What we commit to is the work being done, the measurements being run as described, and the results being reported to you in full, including when they are poor. First movement, where it happens, usually shows six to eight weeks after the work begins. ## What we ask of you Accurate information about your business, and permission to correct your details on directories and listings where we act on your behalf. If information you give us is wrong, the measurement will be wrong with it. ## Payment and cancellation One-off work is invoiced before it starts. Monitoring is billed monthly with no minimum term; cancel at any time and the service runs to the end of the paid period. Prepaid periods are discounted and are refundable pro rata if we stop delivering. ## Your data and your results The measurements, answers and reports we produce for you are yours. You may publish them, hand them to another supplier, or take them with you if you leave. We handle personal data as described in our privacy policy. ## Limits We are liable for the fees you have paid us for the work in question. We are not liable for lost profit or for decisions taken on the basis of a measurement — a measurement describes what assistants said at a point in time, and assistants change. ## Getting in touch Questions about these terms go to hello@ai-sayso.com, and get a plain answer rather than a form letter.