When AI describes a company, it uses the clearest sentences it finds on the site. The list below goes page by page through where those sentences should sit, how they should be written and the markup machines will read.
In What is GEO? we listed the work to be done on a site as headings. This article is the hands-on side: what to write on which page, what to mark up for machines and what not to do. The list was prepared with a corporate site or a service business in mind; e-commerce sites have extra work on the product side.
The door first: robots.txt and access
Everything below only works if bots can get into the site. Things to check:
- There should be no rule in
robots.txtblocking the AI bots that search (OAI-SearchBot,ChatGPT-User,Claude-SearchBot,PerplexityBot). - Whether to allow training bots (
GPTBot,ClaudeBot,Google-Extended) is a matter of preference. Blocking them does not directly prevent visibility in search answers, but it slows down how well models get to know you in general. - Bot protection on a CDN or firewall may be blocking these bots separately; do not relax just because
robots.txtis open. - Important text should sit in the page source as plain HTML, not be loaded later with JavaScript.
Content, page by page
Home page: the first two sentences
AI often takes the answer to “what does this company do?” from the top of the home page. Write a definition, not a slogan: what you do, for whom, where. “Building the future together” answers no question; “We provide CNC machining and tooling for automotive suppliers in the West Midlands” answers many.
About us: identity details
- Registered company name and brand name (both, if they differ, with how they relate).
- Year founded, head office address, area served.
- Team or founder details, and any certifications and memberships.
This page is where information from different sources gets tied to the same company. Contradictions are resolved here.
Service pages: one page per service
Instead of listing every service as bullet points on one page, give each main service its own page. On each page:
- A definition of the service in the first sentence.
- Who it suits (industry, company size, situation).
- The process: how many steps, how long it takes.
- What determines the price; if possible, a range or a starting price.
- 3-5 frequently asked questions with short answers.
Some companies prefer not to give prices; but even a sentence like “the price depends on the floor area and how often the work is done” answers far more questions than “call us for a price”.
FAQ: real questions
Write the questions customers actually ask on the phone and by email, not the ones the sales team would like to hear. The first sentence of each answer should answer the question on its own; the explanation comes after. AI often quotes only that first sentence.
Contact: the same everywhere
Name, address and phone number should be written on the site as text (not inside an image) and match your Google Business Profile, LinkedIn and directory listings letter for letter. “St” versus “Street” is not a problem; a different building number or an old phone number is.
Markup for machines to read
Structured data labels the information on a page using the schema.org vocabulary and is usually added to the page source in JSON-LD format. Visitors do not see it; search engines and AI tools read it. On a corporate site the most useful types are:
| Type | Where | What it describes |
|---|---|---|
Organization | All pages or the home page | Company name, logo, address, contact details, social accounts (sameAs) |
LocalBusiness | Businesses with a physical address | Address, opening hours, service area |
Service | Service pages | Name of the service, description, provider, area |
FAQPage | FAQ sections | Question and answer pairs |
BreadcrumbList | Inner pages | Where the page sits within the site |
Article / BlogPosting | Articles | Headline, author, publication and update dates |
Two rules: the information in the markup must match the information visible on the page, and after adding markup you should check for errors with Google's Rich Results Test or the Schema.org validator.
llms.txt: a short summary file
A plain text file at the root of the site (yourdomain.com/llms.txt). A practical structure:
- The company name as the heading.
- One paragraph on what the company does, for whom and where.
- A list of important pages: on each line, the page name, its address and a one-sentence description.
- Frequently asked basics such as prices, opening hours and service area, if you have them.
The file should be saved as UTF-8, and the server should serve it with the text/plain; charset=utf-8 header; otherwise accented letters, currency symbols and other non-ASCII characters may come out garbled. It is a mistake we have run into in the field more than once.
What not to do
- Leaving important information only in an image, a PDF or a video. If your price list exists only as a PDF, its chances of being read drop.
- Putting information in the markup that is not on the page. Misleading markup can lead to penalties from search engines.
- Leaving outdated information up. A closed branch, an old price, an employee who has left. AI can keep repeating these for years.
- Doing everything once and forgetting it. A theme change can break the schema; a plugin update can break
robots.txt. That is why in GEO Support we check these settings every morning.
Checklist
- AI bots that search are allowed in
robots.txtand on the CDN. - Important text is visible in the page source.
- The first two sentences of the home page define the company.
- The About page has the registered name, address and year founded.
- Every main service has its own page and FAQ.
- Contact details are the same everywhere.
Organization,ServiceandFAQPagemarkup has no errors.llms.txtis up to date and UTF-8.- A measurement was made with the same questions before and after the changes.
The last item is the one most often skipped. To see where your site stands on this list, the free AI visibility check covers bot access, the markup of your company information and whether you are mentioned in the questions you choose, all in a single report.
Frequently asked questions
Do I need to have my site rebuilt to add structured data?
Usually not. JSON-LD markup is added to the head section of the page without changing how it looks. On systems like WordPress it is done with a plugin; on custom sites, with a few lines of code.
Can I appear in AI answers without llms.txt?
Yes, you can. llms.txt is not yet an official standard, and it is unclear which tools read it and to what extent. We recommend adding it because it is easy to prepare, but give priority to bot access and clear page content.
How long should I wait after applying this list?
Technical changes are picked up when bots recrawl the site; that ranges from a few days to a few weeks. To see the effect, ask the same questions again four to eight weeks later and compare the mention rate.
Is there a job that keeps repeating at your company? Tell us about it and we will build the tool that does it.
Tell us what you needOther articles
Can a spreadsheet with thousands of rows be entered into a system in bulk?
In most cases, yes. The hard part is not the entry itself but getting the data into the shape the system expects and verifying that every row was really saved. If the system offers no bulk route, the work goes through the screen, and that depends on the rules of that system.
21 September 2026Read →Why does preparing a quote take so long?
Writing the quote does not take long; the time goes into hunting for the price, copying and correcting an old quote, and rebuilding the layout every time. Once the price list sits in one place and the rules are written down, a quote comes out in minutes.
21 September 2026Read →Which jobs are suited to automation, and which are not?
The job best suited to automation is one that repeats often, whose rules can be written down and whose data is already digital. Work that needs judgement, is full of exceptions or has its data on paper needs another step first.
18 September 2026Read →What if the AI assistant on your site gives a wrong answer?
The risk is real, but manageable. An assistant's wrong answer usually comes from guessing about something it does not know. The solution is to draw a clear line from the start around what the assistant knows and does not know, and to have it hand over to a person when it does not know.
18 September 2026Read →Is ranking in Google the same as being mentioned by AI?
Partly. Both draw on the same foundation, but they do not reward the same thing: Google ranks a page, while AI reads several sources, writes an answer and decides for itself who to mention in it.
17 September 2026Read →Why doesn't ChatGPT mention my company?
The reason is rarely “the AI does not like us”. Usually it is a technical block, scattered company information, or the company not being mentioned anywhere else online. All three can be fixed.
17 September 2026Read →What is GEO? Being visible in AI answers
GEO (Generative Engine Optimization) is the work of making sure that when ChatGPT, Gemini, Perplexity and Google's AI Overviews answer a question, they find your business, describe it correctly and cite it as a source. It does not replace SEO; it builds on top of it.
17 September 2026Read →Will an AI tool learn my company's data?
Short answer: not in a tool running through a commercial API. But that answer depends on how the setup was done, so it is worth knowing what goes where.
16 September 2026Read →When does a spreadsheet stop being enough?
Excel is not a bad tool; it handles most jobs and more. The problem is carrying on with it for months after the point where it stopped being the right fit.
16 September 2026Read →Off-the-shelf software or custom-built?
The right answer depends on how specific the work is to you. Work that everyone does the same way is cheaper off the shelf; work you do your own way is cheaper built for you.
16 September 2026Read →