Blog, page 24 of 28
Notes on getting recommended by AI.
The full archive. Everything we have published on what AI apps read, what they cite, and what moved the numbers.
AI crawlersCan AI search crawlers read your website? A dated test of robots.txt, HTTP access and named crawlersThe named crawler list, the fetch test, the fields to record, and a plain statement of what we have run and what we have not.29 September 2026buyer question topic mapA buyer question topic map: 100 named prompts, scoring rules, and a prioritised content planAll 100 prompts written out and labelled by source type, with a five factor score for deciding what to write first.29 September 2026prove GEO is workingHow to measure GEO without overclaiming: a before and after measurement protocolFour measurement layers, both periods reported in full, and the line where observation stops and causation would begin.29 September 2026which pages get cited by AIWhich website pages earn AI citations? A page type benchmark with URL level resultsEight page types, the URL level counts we can actually account for, and the denominators beside each.29 September 2026where to start with GEOGEO starting point: a 30 day, evidence led website experimentOne frozen prompt set, one month of changes, and the before and after citation counts printed side by side.29 September 2026AI visibility reportAI visibility reports explained: a metric dictionary with worked examplesFourteen metrics defined by their denominators, and how to read a score that has none.29 September 2026AI visibility auditAI search visibility audit: a reproducible 30 prompt method and resultsThirty prompts, three engines, three repeats, and every result reported with its denominator.29 September 2026test ChatGPT recommendationsHow to test whether ChatGPT recommends your business: a 20 prompt benchmarkThe 20 prompts written out, how to run them, and the ChatGPT results we have measured with their dates.29 September 2026Profound vs OtterlyAI visibility tools compared: a reproducible benchmark across 6 toolsA 15 prompt test spec, what each tool must reproduce, and the full cost of running it.29 September 2026alternatives to ProfoundAlternatives to Profound: the numbers and the evidence to require before you buyThe six questions that separate a real visibility measurement from a dashboard, and the counts we can actually show.29 September 2026AI Overviews live searchDoes Google AI Overviews search the web live? What Google says, and a repeatable freshness testSix things live search could mean, a test you can run on ten pages, and why we publish no median.29 September 2026incorrect AI OverviewHow to document and report an incorrect Google AI Overview about your businessA 20 field evidence log, the reporting routes that exist, and why no honest page can quote you a fix rate.29 September 2026AI Overviews auditHow to check whether your business appears in Google AI Overviews: a repeatable auditA two hour audit you can run today, and the honest way to report the share that links your own pages.29 September 2026AI Overviews sourcesWhich sources appear in Google AI Overviews? A dated, clickable source auditHow to count distinct source URLs properly, and why a link count with no query count is meaningless.29 September 2026Google AI OverviewsWhat we can and cannot measure about which businesses Google AI Overviews nameA full capture protocol, the counts we actually hold, and the one rate we refuse to publish.29 September 2026get recommended by ChatGPTHow to get your business recommended by ChatGPT: our own citation counts, the runs behind them, and what to doFour named runs, the counts they produced, the prompt set construction, and the work those counts point at.29 September 2026review sites AI recommendationsDo review sites change AI software recommendations? The 12 week audit method, and what our own captures showHow to count review site citations by named service, and the honest state of our own counts.29 September 2026citable content experimentWhat makes a page citable in AI answers? A preregistered 12 week experiment, published before it is runThe hypothesis, the randomisation, the treatment definition and the analysis plan, written down in advance.29 September 2026why AI cites some pagesWhy AI search cites one page instead of another: the matched page audit method, and what our own audit recordedSeven measurable page features, a matched pair design over 200 queries, and the reasons an engine gave us in writing.29 September 2026AI citation sourcesWhere AI shopping recommendations get their sources: the category audit method, and our own dated findingsEight named source categories, a coding protocol, and the ordering our September 2026 capture actually shows.29 September 2026ChatGPT vs GeminiChatGPT and Gemini product recommendations: the 100 prompt comparison protocol, and what we have run so farHow to measure top pick agreement between two assistants, and an honest account of which part of it we have done.29 September 2026good AI mention rateWhat is a good AI mention rate? A measurement guide without a fake industry benchmarkThere is no published benchmark worth believing. Here are our own raw counts and the method that produced them.29 September 2026OAI-SearchBotHow to check whether ChatGPT search can access a new page: a timestamped crawl test logEligibility is documented, a timetable is not. Here is the log to keep, and what our own record does and does not contain.29 September 2026ask ChatGPT about my companyShould you ask ChatGPT about your own company? A fair test protocolYes, but never only by name. A named prompt and a blind prompt measure two different things.29 September 2026AI mention rateHow to measure whether ChatGPT mentions your company: a repeatable prompt test protocolThree definitions, a frozen prompt set, a log sheet, and rates reported as counts over a denominator.29 September 2026AI recommends competitorWhy an AI answer may recommend a competitor: a diagnostic checklist and a repeatable testSeparate a one-off from a pattern, then check eligibility, search, citations and the prompt itself.29 September 2026ChatGPT recommendationsHow ChatGPT can surface company recommendations: what is documented and what is notThere is no published recommendation formula. Here is the test we ran instead, with the counts and the dates.29 September 2026GEO vs SEOGEO vs SEO: what is different, what overlaps, and what we can actually measureTwo instruments, two denominators, and the matched query method that lets you compare them fairly.29 September 2026generative engine optimisationGenerative engine optimisation: definition, scope, and a reproducible citation testWhat GEO is, what it is not, and the counted result of our own blind runs with the dates attached.29 September 2026Google AI Overviews business mentionCan a business earn a mention in Google AI Overviews? A reproducible 90-day testA named-business rate with its numerator and denominator, the full 90-day study design, and an honest account of what we have run.29 September 2026does Claude search the webDoes Claude search the web for this answer? A surface-by-surface verification guideFive surfaces, five different answers, and how to verify retrieval on the one in front of you.29 September 2026Claude wrong facts about my companyHow to correct wrong company facts in Claude: source diagnosis, reporting steps, and a follow-up logDiagnose which of seven sources caused the error, fix that source, report it, then log four retests without claiming a cause.29 September 2026does Claude know my companyDoes Claude know your company? A repeatable prompt and citation auditFifteen prompts, a fact sheet, and two scores kept apart: accurate, and accurate with a citation behind it.29 September 2026where does Claude get informationClaude's information sources: Anthropic's published statements, web search, and what users can verifyTwo paths in, one of them inspectable, and a claim-level audit for counting citations that actually support the answer.29 September 2026how Claude decides which companies to naDoes Claude have a business-recommendation ranking formula? What Anthropic documents and what we testedNo published formula, and one dated run where the assistant stated its own reasons for the order it chose.29 September 2026get recommended by ClaudeHow to make your business easier for Claude to find and verify: a reproducible 30-day testPublish facts a machine can check, then count how many search-enabled runs cited your domain.29 September 2026does Gemini search the webDoes Gemini search the web for every answer? What Google documents and a repeatable test showsSearch can be used and should not be assumed, the app and the API behave differently, and here is how to check one answer.29 September 2026Gemini wrong facts about my companyHow to correct a wrong Gemini claim about a business: evidence, reporting, and follow-upPreserve the answer, find the page that caused it, fix the source, report it, and log whether it changed.29 September 2026check if Gemini knows my companyA reproducible Gemini visibility check for a business: prompts, scoring, and resultsA fact sheet, a scoring rubric, and a count of correctly identified company facts out of the facts tested.29 September 2026where does Gemini get informationWhere Gemini's answer came from: a source-checking guide with a 100-answer auditHow to tell training memory from a retrieved page, and the audit design for counting displayed source links.29 September 2026how Gemini decides which companies to naWhat Google does and does not publish about Gemini's business recommendationsNo public formula that we can find, a list of what Google does document, and a test you can run instead.29 September 2026Gemini business recommendationHow to test whether Gemini recommends a business: a reproducible 30-day benchmarkThe full benchmark design, and an honest statement of which parts we have run and which we have not.29 September 2026does ChatGPT search the webHow to check whether ChatGPT can find and cite your business: a reproducible testA blind prompt set, a search count read off the interface, and the result of our own run of 78 questions.29 September 2026recruitment agency DelhiHow a recruitment agency in Delhi gets found in AI answersDelhi, Gurugram and Noida are three hiring markets, and an agency that blurs them answers none of the three questions well.29 September 2026recruitment agency MumbaiHow a recruitment agency in Mumbai gets found in AI answersIn Mumbai the sector you hire for matters as much as the city, and the commute matters to the candidate.29 September 2026recruitment agency BengaluruHow a recruitment agency in Bengaluru gets found in AI answersYour office address matters less than the roles and the corridors you actually hire for, and both need to be in writing.29 September 2026dental clinic NagpurHow a dental clinic in Nagpur gets found in AI answersNagpur clinics serve a large region, and a clinic that only names the city gives an assistant half the facts.29 September 2026dental clinic CoimbatoreHow a dental clinic in Coimbatore gets found in AI answersThe city has two names in common use and many well defined localities, and your listings should handle both facts.29 September 2026dental clinic ChandigarhHow a dental clinic in Chandigarh gets found in AI answersA sector number is a real address and a confusing one, because three cities here use the same numbers.29 September 2026dental clinic IndoreHow a dental clinic in Indore gets found in AI answersIndore clinics serve a wide area around the city, and that fact belongs on the website in one sentence.29 September 2026dental clinic LucknowHow a dental clinic in Lucknow gets found in AI answersThe old city and the newer colonies are two different markets, and your listings should not blur them.29 September 2026dental clinic KochiHow a dental clinic in Kochi gets found in AI answersKochi is a set of distinct towns to a patient, and the surrounding towns are not Kochi.29 September 2026dental clinic JaipurHow a dental clinic in Jaipur gets found in AI answersPart of the patient base is local and part travels in, and those two groups need different sentences on your site.29 September 2026dental clinic KolkataHow a dental clinic in Kolkata gets found in AI answersNorth Kolkata, south Kolkata, Salt Lake and New Town are four different answers, and most clinic sites give none of them.29 September 2026dental clinic AhmedabadHow a dental clinic in Ahmedabad gets found in AI answersWhich side of the river you are on changes the answer, and so does whether your site says it.29 September 2026dental clinic ChennaiHow a dental clinic in Chennai gets found in AI answersChennai runs on landmarks and on some very long roads, so your address has to be written the way people actually give it.29 September 2026dental clinic HyderabadHow a dental clinic in Hyderabad gets found in AI answersHyderabad and Secunderabad are two answers to a patient, and the western corridor is a third.29 September 2026dental clinic PuneHow a dental clinic in Pune gets found in AI answersPune and Pimpri Chinchwad are two places to a patient, and appointment timing matters more here than most clinics realise.29 September 2026dental clinic DelhiHow a dental clinic in Delhi gets found in AI answersMany clinics that call themselves Delhi clinics are in the NCR, and that one sentence decides a lot of answers.29 September 2026dental clinic MumbaiHow a dental clinic in Mumbai gets found in AI answersIn Mumbai convenience is measured in stations and walking minutes, and your listings have to say so.29 September 2026
See where you stand before you read another page.
Enter your website for a free check. No account, and it asks real buyer questions rather than your brand name.
Already sure you want this? Join the waitlist. We open on 31 October 2026. It costs an email address. No card, and nothing is charged. Or talk to a person if you have a question first. Or start with the guides, which are step by step.