Articles
AI visibility guides, and the research behind them.
Guides by trade to what a business's own pages give an AI answer, and the audit they draw on. Every figure here carries the method that produced it.
A brand dropped, new hours, a holiday: making the change everywhere at once
A store's change routine: when a brand comes or goes, the hours change or a holiday nears, the same edit lands on the site, profile, feed and markup.
A first pair of progressives: the page to publish before the fitting
The page a first-time progressive wearer reads before buying: who fits and how, the designs by maker's name, bracketed prices, adaptation and remakes.
A narrow lead in machine building, stated as evidence a buyer can check
For a machine builder that leads a narrow niche: applications solved, the installed base by region, standards as named, mentions linked, and a test.
A page each for children's and sports eyewear, and how to test them
What a children's eyewear page and a sports eyewear page state in a store's own terms, with bracketed prices, and a test to run before and after launch.
A painter's monthly panel: job, surface and town, exterior in season
How a painting company reads AI answers each month: named, recommended or first, lines by job, surface and town, exterior lines in season, five columns.
An afternoon with the stock list: checking four assistants' answers
An afternoon's check for an independent store: shoppers' questions on four assistants, each answer read against the shelf and the site, and what to fix.
An optical store in AI answers: which service, which branch, which facts
What named, recommended and first mean for an optical store, a question panel laid out by service, product and branch, and four columns the record adds.
Cabinets, lacquer, wallpaper, historic finishes: a page per specialty
A page per painting specialty: your record as dated counts, products, methods, price basis, and for lead, mold or asbestos only the certifications held.
Capability pages a buyer can compare: specifications in dated HTML tables
Machine envelopes, tolerances, materials, certificates and capacity as dated HTML tables with units, the datasheet beside each, and markup that matches.
Category, brand, place, condition: a monthly reading for one store
What an AI answer names an independent store for: category, brand, place or condition, whether it was recommended or first, and whether the facts held.
Fittings, setups and buying guides: a specialist store's advice in writing
How a specialist store writes its advice down: fittings and setups, whether to book, staff by specialty, dated buying guides, and after-sale work.
From eye exam to repair: the questions of a glasses purchase, in order
The questions of a glasses purchase in order, from the exam and the plan to frames, lenses, pickup and repair, and how a store gathers them at its counter.
In stock near me: stating availability, pickup and the way to check
What a store can state when a shopper asks what is in stock nearby: the lines it always carries, how to check one item, pickup terms and store pages.
Last year's painting jobs by town, beside the AI answers for each
Set last year's painting jobs by town beside AI answers: towns you painted but are missed in, towns you left but are named for, and what to check first.
Late nights and holidays: a pharmacy's hours, published before anyone asks
What a pharmacy states for late nights and holidays: hours by day and date, any on-duty nights as its own facts, what it does after hours, and the markup.
Monthly readings of AI answers for an engineering firm's long sales cycle
A monthly routine for engineering firms: capability and company-name questions scored apart, a dated log of site changes, and the inquiry log beside it.
Named at the right hour: measuring a pharmacy in AI answers
How a pharmacy measures whether assistants name it, need by need, and records the service named, whether the hours were right that day, and the page cited.
Named for the part you can make: AI answers for machine builders and shops
What named, recommended and first mean for a machine shop or builder, a panel of capability questions, and how to check each figure an answer states.
Pickup times, a delivery area, a pharmacist by name: a pharmacy's own facts
The facts a neighborhood pharmacy states as its own: same-day pickup times, the delivery area and its terms, pharmacist hours, services and stock checks.
Planning a move from far away: the pages that answer it, in order
The questions a household asks while planning a move from far away, in the order of the move, and the page on a mover's site that answers each in text.
Privacy on a pharmacy's public pages: reviews, replies, photos, posts
What never appears on a pharmacy's site, posts or review replies, how to reply without confirming anyone was a customer, and what the pages state instead.
Read, dated and signed: a pharmacy's review routine for pages and posts
Who reads a pharmacy's pages, posts and newsletters before they go out, what each sentence is checked against, how sign-off is dated, and what reopens it.
Stating what your pharmacy offers without saying what a medicine does
How a pharmacy states its services, stock and hours without a health claim, where a product's own words come from, and what to do when an answer adds one.
The painting year as a publishing plan: exterior, interior, commercial
A painter's publishing year: when the exterior, interior, cabinet and commercial lines change, each dated, month by month, with the profile in step.
The record of a painting job: surface, prep, paint line and color
What a painting job's record states that a photo cannot: surface and condition, the prep, paint by maker and line, colors, town and month, client words.
The terms a household checks before booking a mover, stated as yours
Quote terms and what changes them, liability cover and its exclusions, licenses as issued, membership and claims: how a mover states each as its own.
Through the moving year: AI readings beside your quote requests
Reading a mover's monthly AI panel through the busy season and month-end peaks, beside quote-form sources, referrals and Google's generative AI report.
Which move were you named for? Reading AI answers by route and move type
An answer names a mover for one move at a time: a type, a route, a date. Defining named, recommended and first, and a panel asked from both ends.
A first-quarter GEO plan for a licensed electrical contractor
An electrical contractor's first GEO quarter: jobs chosen by margin and crew, the license stated once, pages in order, a reading before and after each.
A fulfillment center's public pages and portal: what crawlers can reach
Which of a fulfillment center's pages AI crawlers can reach, the facts a founder asks about as public text, and what stays behind the client login.
A landscaping website AI can read: from photo gallery to project records
Turn a landscaping gallery into project records: town, job, size, season, materials and plants in text, captions on every photo, and markup to match.
A page per cargo, lane or client type, with the record behind it
Why a logistics provider writes one page per cargo, lane or client type it handles repeatedly, what each page states, and how to test it with a panel.
AI visibility for architects: the firm, the principal and the building
An answer about architects can name the firm, a principal or a building. How to state each on your site, credit shared work honestly and measure it.
AI visibility for beauty studios, counted treatment by treatment
A beauty studio is named for a treatment, not in general. Build the question panel from your menu, then read each answer for its condition and stale facts.
AI visibility for car dealerships: measuring a lot that changes every day
What being named means for a dealership, why its panel needs two kinds of question, and the four columns a dealer adds while the lot changes daily.
AI visibility for carpenters: named for which job, and as what kind of shop
Reading AI answers about a carpentry workshop job by job: named, recommended or first, called a maker or a store, and whether lead time and price match.
AI visibility for cleaning companies: commercial and residential apart
Buyers ask for cleaning by service, building and schedule. How to read the answers in two panels, commercial and residential, and what to record.
AI visibility for electricians: named for which job, with which number
What named, recommended and first mean for an electrician, a bracketed panel by job and town, and the record columns that catch a wrong night number.
AI visibility for gyms: named for the right class, on the right terms
What it means for a gym or studio to be named by an AI assistant, and a panel of questions by class, terms, time and place that measures it each month.
AI visibility for hotels: named for which conditions, linked to whose page
What named, recommended and first mean for a hotel, a panel built from trip types and conditions, and the record columns for what an answer linked.
AI visibility for lawyers: the name in the answer and the facts beside it
What named, recommended and first mean for a law firm, a panel by practice area, county and term, and the columns that catch stale facts about it.
AI visibility for management consultancies: named for which problem
What AI visibility means for a consultancy: named, recommended or first for a problem, an industry and a size, and how to read it when work comes later.
AI visibility for marketing agencies: named for what, and on whose work
What AI visibility means for an agency: named for a service, an industry and a budget, on the evidence of its case studies, and the four fields to record.
AI visibility for planners: named for which service, where, and how big
What being named by an assistant means for a wedding or event planner, why it splits by service, place and size, and a question panel to measure it.
AI visibility for plumbing and heating companies: both trades on one panel
What named, recommended and first mean for a plumbing and heating company, a panel by call, season and town, and why heating lines are read by the year.
AI visibility for real estate agents: named, and under which brokerage
AI visibility for an agent: named, recommended or first, a seller and buyer panel by neighborhood, and a record of which brokerage each answer gave.
AI visibility for recruiters: a client reading and a candidate reading
What AI visibility means for an agency asked about by clients and by candidates, and a split panel and record that show which side each answer served.
AI visibility for roofers: by job and town, with storm weeks kept apart
How to read AI answers about a roofing company by job, material and town, with storm-week lines logged apart and every named contractor's address checked.
AI visibility for vets: named for which animal, which service, and when
How a veterinary practice measures whether AI assistants name it, line by line for each animal, service and hour, and checks the hours each answer states.
AI visibility in landscaping: measured by season, service and town
What named, recommended and first mean for a landscaping company, a panel bracketed by season, service and town, and why May is read against last May.
Alternatives and versus pages for SaaS: comparisons a reader can check
How to write an alternatives or versus page for your own market: rivals named fairly, each fact dated to a plan and version, a review date on every table.
An afternoon's check of AI answers against your dealership's lot
An afternoon check for a dealer: eight questions on four assistants, each answer held against the lot that day, and the page behind every wrong answer.
An IT provider's partner designations and certificates, stated as issued
How an IT provider states vendor designations and certificates such as ISO/IEC 27001: the issuer's words, the scope, the dates, and what it does not hold.
Architect nearby: licensed where the site is, not the shortest drive
For an architect, near means licensed where the site is, versed in its approvals and able to reach it. How to state all three so answers can repeat them.
Are you on the bid list? An afternoon's AI check for cleaning companies
An afternoon's check on a facility manager's questions: four assistants, each answer read against your own pages that day, and a fix for each wrong one.
Before the assessment: the heat pump questions a page should answer
What a heat pump page answers before a homeowner books the assessment: sizing, systems, the house, cold weather, noise, price, warranty and incentives.
Burst pipe, sewer backup, no heat: writing the plumbing emergency page
What a plumbing and heating emergency page says: the calls you take tonight, the one gas instruction, the night area and rate, and the paperwork after.
Changelog, docs and trust center: pages that date and prove your claims
SaaS pages that date and prove a claim: a changelog entry per change, open docs with limits, a trust page of what you hold, status history, plan limits.
ChatGPT and a planner's region: OpenAI's crawler rules, then your facts
What OpenAI's crawler page settles for a wedding or event planner, how to check its crawlers can reach you, and the regional facts your pages must state.
Civil ceremony, boho, luxury, elopement: a page for each, and a test
How a wedding or event planner gives each ceremony type, style or format a page of its own facts, and how a question panel shows whether a page helped.
Class timetable, prices and equipment: a gym's pages, written as text
How a gym puts its class timetable in an HTML table, its membership and class prices with what each includes, its equipment by area, and seasonal changes.
Confidential work as evidence: what a consultancy can publish
What a consultancy can publish about work under a confidentiality agreement: anonymized cases with dates and scope, approved names, method and talks.
Costs, scheduling and care: three pages a landscaping customer reads next
The prices, scheduling and aftercare pages a landscaping customer reads after the shortlist, each question a heading answered with your own figures.
Describing the work, not the result: what an advisor's pages can say
What an advisor's pages can describe instead of outcomes, what goes to the compliance officer first, and what to do when an answer promises something.
Direct bookings from AI answers: your own page as the better place to land
When an answer names your hotel, some travelers land on your own page. What that page needs beside the rate, and how to count what the visits did.
Disability cover questions: what an advisor who arranges it states
For an advisor who also arranges disability or income protection cover: the licensed lines, the insurers, how you are paid and what a first review covers.
Emergency electrician inquiries: the page a caller reads in the dark
What an electrician's emergency page tells a caller reading it in the dark, and how its hours, night area and number match the profile and the markup.
Exotics, dentistry, surgery: a page for each service you want asked about
What a veterinary service page states: the animals by name, what is done in house and what is referred, who does it, the equipment and a price range.
Found regionally: how a cleaning company states the area it covers
For a cleaning company whose crews go to the client: one service-area profile, areaServed in the markup, one page of towns, and what changes at the edge.
From Google profile to expert text: what a roofer's listings cannot hold
The pages about a roofer it does not control, from the Business Profile to permit records, how to keep them agreeing, and the pages only its site can hold.
From Google ranking to AI answer: what changes for a gym, and what stays
What changes for a gym when a ranked list becomes an answer with a few names in it, what its pages must state, and the SEO work that still counts.
Glass, post-construction and disinfection: a page for each specialty
Specialty pages for a cleaning company: heights and access for glass, phases and handover for post-construction, and products as named for disinfection.
Guest reviews on your hotel's own pages: what to quote, and the rules
Which guest sentences answer a traveler's condition, the room page each belongs on, how to quote them, and the rules on reuse, markup and replies.
Hallucinated features: find and correct wrong claims about your product
Catch what answers get wrong about your software, from invented features to retired integrations and old prices, then fix the page each claim came from.
How ChatGPT reaches a gym's pages, and why the member app stays shut
What OpenAI's crawler page says about the tokens that reach a gym's site, which pages to keep open, and why the member app stays behind its login.
How many roofing inquiries come from AI? How to count your own
Nobody publishes how many roofing inquiries start with an AI assistant. Count your own: one intake question, the job record, Search Console and one number.
How often does the AI recommend your landscaping business, town by town?
An afternoon's check of AI answers against your route map: the towns you serve, a few you do not, and the page or listing behind each wrong town.
How your advisory firm is paid, stated in words a client can check
How an advisory firm states who pays it, the fee schedule, the minimum and its disclosure documents, in words a client can check before a first meeting.
Law firm marketing and AI: review what you publish, record the rest
Who reads a law firm's page before it goes live, what they check it against, how the approval is dated, and what to do when an AI answer adds a claim.
Lawyers' conduct rules and GEO: state the facts, ask about the claims
Why facts are what an AI answer can repeat about a law firm, seven facts to state and how, and six kinds of statement to take to the bar first.
Making a hotel website machine-readable, field by field
The Hotel and HotelRoom markup a property needs, field by field, what each value must match on the page, and what a booking engine on another host hides.
Measuring AI mentions for hotels: reading the panel through the year
Why a hotel's monthly reading needs a note of the month's events, a comparison with the same month a year on, and the booking numbers kept beside it.
Measuring AI visibility for clients: how an agency grades its own work
An agency that reports AI visibility is grading its own work. What lets a client trust the number: a frozen panel, every answer attached, a dated log.
Measuring mentions for real estate agents: one afternoon, seller and buyer
An afternoon's check for agents: seller and buyer questions on four assistants, each answer read against your own pages that day, each wrong fact traced.
Near-me questions for a beauty studio that is not a storefront
How a studio in a suite, inside another business, at home or on the road states where its door is, which days each treatment runs and the area it covers.
Neighborhood and property-type pages built from an agent's own record
A page per neighborhood you work and per property type you sell, built from your own closings and knowledge, inside MLS rules, and tested with a panel.
Niche or generalist: what a consultancy can state, and how to test it
No measurement of ours shows specialist consultancies get named more. How to state each specialty you can back up, and test the page with your panel.
Niche or grab-bag: a desk page for each specialty your placements back
No measurement of ours says specialist agencies get named more. How to give each desk your record supports a page, with dated counts, and test it.
One property, several businesses: GEO for a hotel restaurant, bar and spa
A hotel with a public restaurant, bar, event space or spa is asked about by different people. Which outlets get a page and a listing, hours, and markup.
Operating figures a buyer can check, each with its period and definition
How a logistics provider publishes cut-offs, receiving windows, throughput and on-time rates so a buyer, and an answer quoting them, can rely on them.
Plumbing and heating service data as one record, marked up to match
How a plumbing and heating company types itself in schema.org, marks up each service and both sets of hours, and keeps the site and profile in step.
Proof a beauty client can check before booking, and where it goes
What a studio can show a client before a first booking: license lookups, dated training, product lines, written policies, aftercare, reviews and mentions.
Reading AI answers about your law firm, practice area by county Start with one afternoon
An afternoon's check of how four assistants answer for a law firm, by practice area and county, each answer read against the firm's own pages that day.
References and a list of buildings a facility manager can check
Evidence a cleaning company can offer a buyer: a list of buildings by type and size, contract lengths as its own counts, references on request, reviews.
Reviews, real weddings and vendor credits: a planner's evidence in text
Which reviews a planner should quote, how to write up real weddings and events with permission, what vendor credits are, and what stays out of markup.
Selling GEO as an agency service without promising a place in the answer
How an agency adds GEO: its own dated readings first, the audit, fixes and monthly reading as separate lines, and a contract that never promises placement.
Situation pages for an advisory firm, and a test of what they change
How an advisor turns an all-purpose services list into a page per client situation, with its own counts, process, fee and limits, and then tests the page.
Solar and the roof: what a roofer's page answers before panels go on
A roofer's solar page answers: replacing the roof first, who installs the panels, how mounts are sealed and warranted, removal and reinstall, and prices.
Spring to winter: a landscaper's publishing calendar, month by month
When each season's page goes up, the dated booking line it carries, what to take down as a season ends, and how the Business Profile keeps in step.
Temp staffing: your licenses and terms, stated for clients and workers
What a temporary staffing agency publishes about itself: each license with its number and the authority's lookup, and its terms for clients and workers.
Temperature-controlled freight: what to publish for a buyer who audits
What a temperature-controlled carrier or warehouse states for a buyer who will audit it: ranges, monitoring, records, certificates, lanes and excursions.
The after-hours page for a vet practice: who answers, and where to go
What a vet practice's after-hours page states: when it sees emergencies itself, the hospital it refers to, the call first, the fee basis and the markup.
The EV charger page: what a buyer asks before booking an install
What an electrician's EV charger page has to answer before a buyer books: who supplies the charger, the panel check, permit, install time and price range.
The storm damage page: what a roofer writes before the storm arrives
What a roofer's storm page says before hail or wind: the first visit, an honest waitlist, where you are based, the records you hand over, and the hours.
The used-car page as a record: one vehicle, every fact a buyer checks
What a used-vehicle page states so an answer can be built from it: the facts a buyer checks, price and fees, the sold notice, and the Car markup.
Three places an advisor's facts live, and which ones AI crawlers reach
Which of an advisor's pages OpenAI's and Perplexity's crawlers can reach: the public site, the client portal behind a login, and the regulator's register.
Timber frames, historic buildings, energy standards: a shop's own record
How a carpentry shop states a specialist record for timber frames, historic buildings and energy-standard builds: projects, approvals and certifications.
Trust without an interview: what an owner can check before a first visit
What an owner can check about a vet practice before a first visit: licenses, accreditation, the team, estimates, policies, reviews and end-of-life care.
What a made-to-measure workshop can state that a flat-pack store cannot
What a made-to-measure workshop can publish and flat-pack cannot: its process, woods and joints, lead time and price from real jobs, and its aftercare.
What building owners ask assistants about architects, stage by stage
Owners ask about an architect in the order of the project: feasibility, choice, fees, time, construction. Examples, and where each answer belongs.
What ChatGPT and Perplexity can reach on a recruiting agency's site
Which of an agency's pages OpenAI's and Perplexity's crawlers can reach: the public site, job listings on an applicant-tracking host, and the portal.
What clients and candidates ask about recruiters, in the order they ask
The questions hiring managers and candidates bring to a recruiter, in the order each side asks them: examples to collect and answer on each side's page.
What clients ask about a beauty treatment, in the order they ask it
A client's questions follow a treatment: before booking, before the visit, after it, between visits. How a studio collects and answers them in writing.
What couples and event clients ask, stage by stage, and where you hear it
The questions couples and companies ask before hiring a planner, stage by stage, where you already hear them, how to collect them, and two answer blocks.
What customers ask about electrical work, from booking to inspection
An electrician's customer questions in the order of a job, how to collect them without personal details, and the one answer to do-it-yourself questions.
What homeowners ask a carpenter, and the page each answer belongs on
A commission's questions, from 'can you make this' to adjusting a door after fitting, where to collect them, and which workshop page answers each.
What homeowners ask before choosing a listing agent, and what to publish
The questions a homeowner asks before choosing a listing agent, and the evidence to publish: dated sales, a marketing plan, seller reviews and your terms.
What investors ask before, during and after hiring an advisor
The questions a client asks about an advisor, from choosing one to leaving, where each answer goes on the site, and how to collect your clients' questions.
What people ask AI before buying a car, in the order they buy
A car buyer asks in the order of the purchase. Which questions a dealer can answer, the page each belongs on, and how to collect your own from the BDC.
What pet owners ask a practice, from the first visit to the last
The questions owners bring a vet practice over an animal's life, which the site answers in writing for every owner, and which get one sentence: call us.
When your own bot protection turns the AI crawlers away: a check for MSPs
An MSP's own firewall or CDN challenge can turn OAI-SearchBot and PerplexityBot away while robots.txt allows them. How to let them in, and test it.
Which lanes AI answers name you for, and how current their facts are
How a logistics provider reads AI answers: named, recommended or first for a lane and a cargo, and whether what the answer says about capacity is current.
Women-only hours, training after injury, a boutique format: a page each
How a gym states an offering it wants to be asked about on a page of its own, in its own policies: women-only hours, training after injury, boutique.
A café menu machines can read: milks, pastries and the seasonal board
How to publish a café menu as text: drinks and sizes, milks and extras, the bakery case, your own dietary labels, seasonal items, and Menu markup to copy.
A dental practice, ChatGPT and Gemini: which crawlers reach the site, and what they find there
Which OpenAI and Google crawlers reach a dental practice's site, what each is documented to do, what they find there, and how to check.
A GEO strategy for an online shop: the first month's four decisions
The first month of GEO for an online shop: which product lines and questions lead, who owns the product data, what is off the list, and how it is judged.
A good hairdresser near me: the place and time facts a salon must state
What a salon's site must state for a near-me question: one address, the neighborhood names clients use, parking and transit, hours by day, and walk-ins.
A portfolio a crawler can read: captions, alt text and a page per shoot
What a crawler gets from a gallery of photographs, and how a page per wedding or project, visible captions, alt text and file names give it words to read.
A repair shop's Google Business Profile: categories, services and hours
The Business Profile fields that matter for a repair shop: categories, services, holiday hours, attributes and service area, each matched to the site.
AI search engines at a glance: who fetches your pages, and how each answers
ChatGPT, Claude, Gemini, Google AI Overviews, Perplexity and Copilot side by side: which crawler fetches your pages for each, and what it shows the buyer.
AI visibility for cafés: what being named means, and how to count it
What it means for a café to be named by ChatGPT, Claude, Gemini or Perplexity, and how to measure it with a fixed panel of the questions people ask.
AI visibility for dentists: named for the filter the patient added
What being named by an assistant means for a dental practice, and how to measure it with the filters patients add: new patients, plans, urgency.
AI visibility for doctors: what being named means, and how to measure it
What it means for a medical practice to be named by an AI assistant, how to measure it with a fixed panel of patient questions, and what a baseline shows.
AI visibility for online shops: named as the shop, the brand or the product
What being named in an AI answer means for an online shop, its brand and its products, and how to build a question panel by product need and condition.
AI visibility for photographers: named for which work, and where
What being named by an assistant means for a photographer, why it differs by genre and by place, and a panel of client questions to measure it with.
AI visibility for physiotherapy practices: what being named means
What it means for a physiotherapy practice to be named by an AI assistant, and how to build a panel of patient questions by condition, treatment and place.
AI visibility for restaurants: named for tonight, with the right hours
What being named by an assistant means for a restaurant, how to measure it with questions of cuisine, constraint, time and place, and what to check.
AI visibility for salons: what being named means and how to measure it
What it means for a salon to be named by ChatGPT, Claude, Gemini or Perplexity, the client questions a salon's panel holds, and how to read a baseline.
AI visibility for tax advisors: being named for the client's situation
What it means for an assistant to name your tax firm, and for which situation, and how to build a panel of client questions by situation, place and season.
Answer engine optimization (AEO): where it came from, and what it asks of a page now
AEO began with featured snippets and voice answers, the one passage an engine lifted from a page. What carried over to AI answers, and what to skip.
Ask the AI about your café by name, then check every fact it gives
The questions to ask each assistant about your café by name, the facts to check in every answer, where a wrong one survives, and how to correct it.
Backlinks versus mentions: what each one tells a search engine and an assistant
A backlink is a pointer a search engine counts; a mention is a statement an assistant reads. What each does, which systems use which, and what to earn first.
Balayage, curly cuts, extensions: what each specialty page must say
How to write a page for each salon specialty: what it is, who does it, how long it takes, the price, the consultation, aftercare, and photos with captions.
Bathroom renovation inquiries: what a homeowner asks an assistant first
What homeowners ask an assistant before hiring a plumber for a bathroom renovation, what the site has to state, and why the emergency call is a separate page.
Beyond the portfolio: prices, travel, terms and delivery as text
The pages that decide a photography booking, as answers: packages and prices, dates, travel, contract and retainer, delivery, and who holds the camera.
Brand when an assistant does the describing: the sentence it repeats about you
When an assistant names your business it adds a sentence about you. Where it comes from, how to write the one you want repeated, and how to check it.
Broker or comparison site: what a broker's own pages have to state
A comparison site and a broker meet different needs. What a broker's pages must state about advice, placement and claims help for an answer to say so.
Building a GEO strategy in 30 days: seven decisions for the first month, written on one page
A GEO strategy is a set of decisions, not a task list: which questions it is for, which services lead, who owns what, what to skip, and how to judge it.
Building an entity: how a machine learns that your business is one thing
An entity is a business a machine can identify across pages: how @id, sameAs and one node per thing turn scattered facts into one business to name.
Café reviews and listings: the drink people name, and one set of hours
Which café reviews to quote on your site as text, why one that names a drink or a seat says more, and how to keep hours the same on every listing.
Calculating the ROI of GEO: a break-even sum you can do with your own numbers
The ROI of GEO is a break-even question: how many extra customers a year would pay for the work. The sum, the numbers only you have, and the limits.
Carpenters in ChatGPT and Gemini answers: what the shop's own site has to say
What homeowners ask ChatGPT and Gemini about carpenters, what your site has to state in text, which schema type to use, and what to fix first.
ChatGPT and a physio practice: what OpenAI's crawler page settles
What OpenAI's crawler page says about the tokens that reach a physio practice's site, what stays behind the booking login, and how to check access.
ChatGPT and your repair shop: what gets fetched, and what a page must say
What OpenAI's crawler page says about the tokens for ChatGPT search, user fetches and training, and what a repair shop's service pages must state.
Collecting the questions patients ask, and turning them into answer pages
Where a medical practice finds the questions patients ask, how to record them without patient details, which become answer pages, and a template for one.
Competitor analysis in AI answers: find who is named instead of you, then read their pages as a crawler does
Competitor analysis in AI answers starts from the answers: who the assistants name for your buyers' questions, what they cite, and what those pages state.
Dealer service pages: booking, warranty, recalls and the makes you service
What a dealership's service department must state apart from its sales pages: booking, warranty and recall work, makes serviced, hours and its own listing.
Developing a GEO content strategy: which pages to write, in what order, and how to keep them true
A GEO content strategy for a small business: the pages worth writing, where their topics come from, the order to write them in, and how to keep them true.
Directories and review portals: which listings matter, what each needs, and the facts that must match
Which directories and review portals a small business should be listed on, what each listing needs, and why the name, address, phone and hours must match.
Documenting reference projects so they count as evidence, not just photos
A reference project is evidence only once its facts are in text: the job, the place, the date, the size, what was hard. How to write one up, with photos.
Does the AI recommend your dental practice? A 30-minute check, and what to do with the result
A 30-minute check a dental practice can run itself: six patient questions, four assistants, the facts to verify, and where each result points.
FAQ schema after Google's change: what it still does, and how to write it
Google limited FAQ rich results in 2023 and stopped showing them in May 2026. What FAQPage markup still does, what it must match, and a block to copy.
GEO for B2B companies: being on the shortlist a buyer asks an assistant for
How a B2B company gets onto the shortlist a buyer asks an assistant for: the pages to state, the proof to publish, the mentions that matter, and measuring.
GEO for e-commerce: where a small shop has room to be named
Where a small online shop can be named in AI answers: questions about its own name, its own brand and narrow needs. What to do first, and what to leave.
GEO instead of pure SEO? A decision by how your customers buy, and the work to do first either way
Should a small business do GEO instead of SEO? Neither alone. How to decide where the next hour goes by how customers buy, and what to do first either way.
GEO KPIs: the seven numbers worth reporting, and the ones to ignore
The numbers that show whether AI visibility work is working: named, recommended and first rates, share of voice, accuracy, access and the four site facts.
GEO reporting for management: the one-page report, and the decisions it should ask for
A GEO report for management fits on one page: a verdict in words, the rates and their trend, what the assistants got wrong, and the decisions to ask for.
GEO vs SEO, task by task: which work helps search, which helps AI answers, and which helps both
GEO and SEO share most of their work. A task-by-task list of what helps search rankings, what helps being named in AI answers, and what helps both at once.
Getting onto the AI shortlist: from mentioned, to recommended, to first
An assistant's answer is a shortlist of two or three names: what separates being mentioned from recommended and from first, and what moves each.
Google AI Overviews: what Google says helps a page get cited, and what does not
What Google's documentation says decides which pages an AI Overview cites, what it says you do not need, and what our crawl found on the pages it picks from.
How ChatGPT and Gemini get to a driving school's pages
Which OpenAI and Google crawlers fetch a driving school's pages for ChatGPT and Gemini, what each token controls, and how to check what they can read.
How to get your salon recommended: the fix order, start to finish
The order to fix a salon's site for AI assistants: a crawler check, one identity, the menu as text, stylists, hours, markup, reviews, listings, a reading.
How to measure AI visibility, and know when an assistant changed its answer
How to measure whether ChatGPT, Claude, Gemini and Perplexity name your business: a fixed question panel, repeated runs, a score, and how to read a change.
Implants, whitening and cleanings: what a dental practice's treatment pages have to say
What a dental practice's pages for implants, whitening and cleanings should state: who does it, what it involves, how many visits, and what it costs.
Is your shop named when drivers ask an assistant? How to check, job by job
What being named by ChatGPT, Claude, Gemini or Perplexity means for a repair shop, and how to check it with a fixed panel of drivers' questions.
Laptop-friendly café: answering the wifi, outlet and time-limit questions
How a café answers the working questions on its own site: wifi, outlets, laptop hours, noise, seating and how long people may stay, limits included.
Large language models, explained for a business owner: what they are, how they answer, and why they get things wrong
What a large language model is, how an assistant turns one into an answer about a business, why answers vary and go wrong, and what that means for a site.
Local AI search with more than one location: one business, several places, stated so an answer can tell them apart
How a business with several locations states one brand and several places so an assistant names the right one: location pages, listings and markup.
Local AI visibility for a medical practice: getting the place right
How a medical practice states where it is: one address format, suites in medical buildings, several offices, hospital affiliations, directories that agree.
Local AI visibility: how a local answer is built and what to fix first
What local AI visibility is, how an assistant builds an answer about a business nearby, what 9,483 local sites got wrong, and four fixes in order.
Making a restaurant menu readable for AI: every dish, its labels and its price, in text
How a restaurant puts its menus online as text an assistant can read: each dish with its label and price, several menus, allergens and weekly changes.
Manual therapy, jaw (TMJ) treatment, lymphatic drainage: a page for each
What each physiotherapy treatment page should state, worked through for manual therapy, jaw (TMJ) treatment and lymphatic drainage, and what to leave out.
Measuring AI mentions: the ones in answers, and the ones on the pages answers are built from
AI mentions come in two kinds: your name in an assistant's answer, and your name on the pages answers draw from. How to count each, and what they tell you.
Measuring AI recommendations for a shop without clean click tracking
How an online shop measures AI recommendations when clicks go missing: a product-question panel, referrals as a floor, an order question, and cohorts.
Mentions and digital PR: what a mention does for AI visibility, and how to earn one
A mention is your business named on a page you do not control. Why an assistant reads them, where a small business earns them honestly, and what to avoid.
One page per genre: what it lets a photographer state, and how to test it
Why each genre you want booked needs a page of its own, what that page can state that a homepage cannot, and how a question panel shows if it worked.
One page per situation: making a tax firm's specialties visible
How a tax firm gives each specialty or client situation a page of its own, what the page must state, what it makes possible, and how to tell if it worked.
Optimizing your own website for AI: the technical floor, sitewide and page by page
The technical floor a small business site needs before AI answers can use it: sitewide settings, then what each page type must return, state and mark up.
Pass rates and reviews: how a driving school states them honestly
When a driving school should publish a pass rate and what it must carry, reviews that name the instructor and the test, and Google's rule on review markup.
Planning a GEO budget: the hours, the money, and who does what
How a small business plans GEO in hours and money: the one-off work and the monthly work, who can do each part, what hiring costs, and how to judge it.
Practice-area pages for a law firm: the matter, the place and the fee, stated so an answer can use them
What a law firm's site has to state, matter by matter, so an answer to a legal question has something of the firm's to use, and what not to claim.
Prices, packages and holiday courses: a driving school's prices page
What a driving school's prices page should state: packages, the price per lesson, intensive and holiday courses, what is included, and cancellation terms.
Product data for AI: a product page that answers every condition
How to write a product page an assistant can quote: the attributes buyers filter on as text, one record behind page and feed, variants, stock, size, fit.
Reddit, forums and communities: taking part honestly, and what it can do for AI answers
Assistants can cite community threads. How a small business takes part honestly on Reddit and forums, what to say and never do, and how to check it helped.
Restaurant data across portals: one set of hours, one menu, one price range, everywhere a guest looks
A restaurant's hours, menu and details live on its site, its map listing, reservation platforms and delivery apps. How to keep them the same everywhere.
Reviews as an AI signal: why an assistant sees none of yours, and where to put them
64.7% of 9,483 small business sites showed a crawler no review signal. How an assistant reads reviews, where to put them on the page, and the markup rules.
Schema.org for online shops: Product, Offer, shipping and returns markup
Product, Offer, shippingDetails and MerchantReturnPolicy markup for an online shop, as Google documents it, in a bracketed template matching the page.
Self-pay patients and AI search: stating prices and what they cover
Why self-pay patients ask about price and scope, how a practice states self-pay prices or ranges, which services are self-pay, and the markup behind it.
SEO vs GEO for someone who already does SEO: what carries over, what changes, and what to stop assuming
For anyone who already does SEO: which skills carry over to GEO unchanged, which change shape, and which ranking assumptions do not hold in AI answers.
Setting up llms.txt: what the file is, what it is not, and where it ranks
What llms.txt is (a proposal, not a standard), what to put in one, where it goes and how to check it, what 9,483 sites did, and the work to do before it.
Setting up prompt monitoring: what to ask, how often, and what to record
What prompt monitoring is, how to build a fixed panel of buyer questions, why one run is a coin flip, how to record each answer, and what the score is for.
Setting up schema.org markup correctly: where it goes, how it is generated, and how to keep it true
How to set up schema.org markup on a small business site: where the JSON-LD goes, builder and plugin traps, rendered or injected, testing, and upkeep.
Share of voice in AI answers: how to count it, and how to grow it
Share of voice in AI answers is your part of every business name the assistants give for your buyers' questions. How to count it, weight it, and grow it.
Specialty pages for a repair shop: a make, EVs, transmissions, fleets
A page per specialty (a make, EVs, transmissions, classic cars, fleets) lets a repair shop answer a narrower question. How to build one, and test it.
Spotting AI traffic in Google Analytics, and what it can and cannot tell you
Finding visits from ChatGPT, Perplexity, Gemini, Copilot and Claude in Google Analytics, why Google's AI results hide in organic search, and the gaps.
Structured data for AI: what to mark up, and what only describes your website
What schema.org markup does for an assistant, why most small business sites carry the kind that describes the website, not the business, and a block to copy.
Structuring content for AI: the heading structure that gets quoted
Which heading structure works for AI-readable content: a heading that is the question, the answer in 40 to 80 words, one idea per section, in real HTML.
Tax firms and ChatGPT: open the public pages, keep the portal shut
Which OpenAI crawlers reach a tax firm's site, why the client portal stays behind its login, how to test access from outside, and what public pages need.
The 10 most common GEO mistakes, and how to avoid each one
Ten common GEO mistakes, most counted in our audit of 9,483 small business sites, from a bot wall nobody chose to a promised placement, with each fix.
The class of 2027: how a driving school reaches next year's new drivers
How a driving school reaches next year's learners: find the season in its own bookings, publish next season's dates and prices early, write for the parent.
The GEO audit: what a real one contains, and how to tell it from a site scan
A GEO audit measures what the assistants say about you, then traces why: the parts a real one contains, how to tell it from a site scan, and what next.
The PDF trap: when your price list is a file, an assistant cannot quote it
Menus, price lists and brochures kept as PDFs, images or script-built pages are invisible to a fetcher that reads HTML. What the crawl found, and the fix.
The questions photography clients type, and where you already hear them
The questions couples, families and businesses ask about a photographer, where you already hear them, how to collect them, and two answer blocks to copy.
The questions physio patients ask, collected and answered in writing
Where a physio practice already hears patients' questions, how to collect them without patient details, and two written answers: referrals and sessions.
Thought leadership a machine can count: publishing your own numbers, with the method attached
Opinion is hard to cite; a number with its period, count and method is not. How a business publishes its own figures so they survive being quoted.
Tracking AI visibility: a routine for the week, the month and the quarter, and what to ask of a tool
Tracking AI visibility is a routine, not a dashboard: what to check each week, month and quarter, the change log beside it, and what to ask of a tool.
Treatment pages that survive being quoted: how to build one
How to build a physiotherapy treatment page so a passage lifted from it stays true and names the practice, and how to see which pages answers cite.
Visibility on Perplexity: what its crawlers fetch, and what a page has to give them
Perplexity shows sources on every reply. What its two crawler tokens do, why blocking one is not privacy, and what a page must state to be cited.
Visible within the rules: what an insurance broker's pages can state
How a broker stays visible to AI assistants inside the advertising and disclosure rules where it is licensed: the facts to state, what to ask counsel.
Wedding photographers and ChatGPT: the crawlers, and what couples ask
Which OpenAI crawlers read a wedding photographer's site, how to check they can, and the dates, venues, packages, travel and delivery a couple asks about.
What AI visibility means for a driving school, and how to measure it
What it means for a driving school to be named by an AI assistant, and how to measure it with a fixed panel of learner and parent questions.
What being named means for an insurance broker, line by line
What it means for an assistant to name your brokerage, for which line of insurance and situation, and how to build a buyer question panel that shows it.
What ChatGPT can read on a medical practice's site, and what it cannot
Which OpenAI crawlers reach a medical practice's site, what they can and cannot fetch, from portals to scripts, what a page needs, and how to check.
What clients ask an assistant before choosing a tax advisor
What a small-business owner asks an assistant before choosing a tax advisor, what your site has to state to be named, and what our crawl found on these sites.
What is generative engine optimization? Origin, meaning and limits
Generative engine optimization: the term's origin in Aggarwal et al. (KDD 2024), what the paper measured, what GEO work is, and what it cannot promise.
What IT decision-makers ask an assistant before they shortlist an MSP
What an IT buyer asks an assistant before shortlisting an MSP, what your site has to state to be named, and why your own bot protection may be in the way.
What learners and parents ask an assistant about driving lessons
The questions learners and parents type about driving lessons, where a school already hears them, and two answers written for a school's own pages.
What patients ask an assistant before choosing a practice
What patients ask an assistant before choosing a doctor, what decides whether a practice gets named, and the facts the practice's own site has to state.
What travelers ask an assistant before they book a hotel
The questions a traveler puts to an assistant before booking, what decides whether your hotel gets named, and what your own site has to state in plain text.
What your Google reviews mean for your salon, and which ones to use
Why salon reviews that name a stylist and a service are the useful ones, how to put them on your pages, how to ask and reply, and Google's markup rule.
When AI gets facts about your business wrong: why, and what fixes it
Why an assistant states wrong facts about a business, how to find out what it says, how to correct the sources it reads, and what no operator lets you do.
Where search is heading: what is already true, what is still uncertain, and what to do either way
What is already documented about search and AI answers, what nobody can know yet, and the work that pays whichever way it goes, for a small business.
Who gets recommended when a founder asks an assistant for a consultant
What a founder asks an assistant before hiring a consultancy, why a stated niche gets matched, and what the site has to say about its people and work.
Why GEO matters now: the case without the borrowed statistics
The case for GEO now, from what can be checked: answers name few businesses, most sites omit the facts, the work is shared with SEO, and it is measurable.
Why isn't my SaaS showing up in ChatGPT? What the assistant can read
Why a SaaS product is missing from ChatGPT's answers: what an assistant can read about a category and a product, and what to publish so it can be matched.
Wikipedia and Wikidata for a small business: what the rules allow, and what to do instead
What Wikipedia's notability and conflict-of-interest rules mean for a small business, what a Wikidata item can hold, and the honest order to build an entity.
Win locally as a mover: what buyers ask an assistant and what to state
What people ask an assistant about movers, what your site has to state to be named for the towns you serve, the MovingCompany type, and the fix order.
Winning clients through ChatGPT: how a question becomes a shortlist
How a buyer's question to an assistant becomes a shortlist, the four things your site has to state, how to measure it, and a realistic first 90 days.
Writing to be cited: what makes a passage quotable by an AI assistant
What makes a passage quotable by an assistant, the answer-block shape with a service-page template, what 9,483 sites lacked, and what each page must state.
Your Google Business Profile and AI visibility: what the listing has to agree with
A Google Business Profile is one more page that describes your business. What Google's guidelines require in it, what your site must say identically, and why.
Zero-click search: what it is, who loses the visit, and what to measure
A zero-click search ends without a visit: the answer was served in place. Which pages keep visits, how to write for quote and click, and what to measure.
robots.txt for AI crawlers: what to allow, and what each token controls
Which AI crawler tokens to allow in robots.txt, what each one controls, why blocking GPTBot does not remove you from ChatGPT, and a file you can copy.
Why ChatGPT does not know your company, and what fixes each cause
An assistant can only name a business it can read and pin down. Six reasons it cannot, what our crawl of 9,483 sites found behind each, and the fix for each.
What 9,483 small business websites actually serve to an AI crawler
Two thirds of small business websites give an AI assistant nothing to quote. What 9,483 sites across 40 trades and 90 metros serve, and what to fix first.
What gets published here
Two kinds of article: research this site ran itself, reported with its method, and guides that turn it into what to change, most of them written for one trade. Both keep one rule for figures: a percentage is one our own audit published, and the page that quotes it links the audit or its method, or it is not printed. Anything taken from outside is named with its source beside it, so you can disagree with the source rather than with us.
How to cite what you find here
Take the numbers; they are not ours to withhold. When quoting one, carry the sample size and the date with it — “9,483 small business sites, September 2026” — because the figure means nothing without them and it will be out of date soon enough. Questions and corrections go to hello@gaash.ai.