How to Structure Your About Page for AI Search

Your About page is doing a job you probably never assigned it.

When someone asks ChatGPT "who runs example.com," when Perplexity is asked whether a company is legitimate, or when Google's systems try to work out if a real organization stands behind your content, the About page is the page they reach for first. It is the one page on your whole site whose entire purpose is to answer a single question: who are you.

Most About pages answer that question with a story. "It started with a laptop and a lot of coffee." That is fine for a human reader who is already interested. It is close to useless for a machine that has been asked a direct question and needs a fact to hand back. The gap between those two things, a page a person enjoys and a page a machine can actually use, is what this post is about.

Why AI systems land on your About page

There are four reasons the About page carries more weight than any other page for identity, and they stack.

It is linked from everywhere. Your About page usually sits in the main navigation or the footer, which means it is linked from every other page on the site. A crawler following your own internal links reaches it early and often. Pages that are linked from one place get visited once. The About page gets visited as a matter of course.

It is where identity questions resolve. "About [brand]," "who founded [brand]," "who owns [brand]," "is [brand] a real company," "where is [brand] based." These are real queries with real intent behind them, and the page most likely to be retrieved and quoted in answer to them is the one with "about" in the URL.

Google tells you to put your identity data there. From Google's Organization structured data guide, on where the markup belongs:

We recommend placing this information on your home page, or a single page that describes your organization, for example the about us page.

Human reviewers use it, and machines want the same evidence. Google's guidelines for the people who manually rate search quality tell them to find out who is responsible for a website, and to use the About page and the contact page to do it. A page thin enough to fail that check when a person reads it is thin enough to fail when a model reads it. The evidence a reviewer looks for and the evidence a model looks for are the same evidence.

One more point underneath all of this. Your About page is prose, and AI systems read prose directly. Your structured data matters, and there is a section on it below, but the sentences on the page are read as plain text too. When the prose and the markup disagree, the page looks less reliable, not more. A half-true claim is treated as worse than an honest gap, which is a principle that runs through the whole Understand pillar of an AI readiness score.

The 30-second extraction test

Open your About page and read it once, at a normal pace. Then, without looking back, try to write down these eight facts:

1. Your exact name, and your legal name if it differs from your trading name.
2. What you do, in one plain sentence, with no adjectives in it.
3. When you started.
4. Who runs it: real names, real roles.
5. Where you are based, and separately, where you operate. A firm in one city can serve a whole country, and those are two different facts.
6. Who you serve.
7. Why you are credible: history, size, credentials, notable clients or work.
8. How to reach a human.

Most About pages deliver four of these at best, and usually the four that matter least. The founding year and the origin story are there. The plain description, the named people, the service area, and the clear "here is who we are" sentence are missing, buried, or written in language too vague to extract.

A model reading your page has less patience than you just did, and no prior interest in your company. If you had to hesitate on any of the eight, so will it.

Write the prose so the facts come out

The fix is not to strip the personality out of the page. It is to make sure the extractable version of each fact is present, near the top, in plain words. The story can stay.

Lead with a definition sentence. The first paragraph, ideally the first sentence, should read like a dictionary entry for your company. "Bright Coffee Roasters is a specialty coffee roaster in Leeds that supplies single-origin beans to independent cafes across the UK and Ireland." That is the sentence a model quotes back when someone asks what you do. "We believe great coffee brings people together" is a nice thing to believe and it answers nothing.

Here is the same opening written both ways. The version most About pages use:

Great coffee is about more than caffeine. It is about connection, craft, and the people who make it happen. Our journey began in a tiny kitchen with a borrowed roaster and a stubborn belief that things could be done better.

And the version a machine can use:

Bright Coffee Roasters is a specialty coffee roaster based in Leeds, founded in 2016 by Maria Santos. We buy green coffee directly from farms in Ethiopia, Colombia and Guatemala, roast it in small batches, and supply it to around 120 independent cafes across the UK and Ireland.

The second one still has a story to tell further down. It just leads with the facts, and every one of them, the type of business, the location, the founder, the founding year, the sourcing, the customers, the scale, is a fact an AI system can lift and hand back correctly.

Shape your headings like the questions people ask. "Who runs Bright Coffee Roasters?" is a better heading than "Our Team." "Where we roast and who we supply" beats "Coverage." A model scanning your headings can map a question-shaped heading straight to the question it was asked. A one-word label makes it guess.

Name people, give them roles, and link their bios. "Founded in 2016 by Maria Santos, formerly head roaster at Kudu Coffee" is worth more than a wall of smiling faces under the word "team." A named, linkable person is a strong identity signal, and each of those people can be modelled as an entity in their own right, the same work covered in the guide to Person and author schema.

State facts as facts. Dates, places, and numbers, written plainly. "Based in Leeds, serving the UK and Ireland since 2016" rather than "proudly serving customers far and wide for nearly a decade." The second version makes a machine do arithmetic and geography to recover something you could have just said.

Separate where you are from where you work. These get collapsed into one line constantly, and they are different questions with different answers. A law firm with one office in Manchester might take clients across England and Wales. A plumber based at a yard in Leeds might cover a 20-mile radius and never see a customer at that address. When someone asks an AI assistant "is there a good [type of business] near me," the service area is the fact that decides whether you are a candidate. Say it directly: the city you operate from, and the region, radius, or list of areas you actually serve.

Keep the origin story, and put it after the facts. A reader who wants the story will scroll for it. A machine that wants the facts should not have to. Lead with what you are and what you do now, then tell the story of how you got here. Both audiences are served, in the right order.

Do not lock identity into a format a crawler cannot read. A founder's letter saved as a JPEG. A team section that only appears after a click or loads in from a script. A company timeline drawn as an infographic with the dates baked into the picture. An "our story" video with no transcript. If a fact exists only inside an image or a script, assume that for an AI system it does not exist at all.

The schema layer, and where it belongs

Two page-level types apply to an About page, and choosing between them matters.

AboutPage is the general one. Schema.org defines it simply as a "Web page type: About page." It sits under WebPage, it has no rich result of its own in Google search, and its job is to label the page for what it is and connect it, through the mainEntity property, to the organization the page describes. That connection is the point: it says "this page is the canonical description of that entity."

ProfilePage is the narrower one, and it is narrower than most people assume. From Google's Profile Page documentation:

ProfilePage markup is designed for any site where creators (either people or organizations) share first-hand perspectives.

Google's own examples of a valid ProfilePage are an author page, an "about me" page on a blog, a forum user profile, an employee page. Its example of an invalid one is a shop's homepage. So the rule of thumb: if your About page is about a company with staff, use AboutPage with an Organization as the mainEntity. If your About page is essentially about one person, a solo consultant, a personal brand, a creator, then ProfilePage wrapping a Person is the honest fit.

The Organization node itself should have one home, either the homepage or the About page, as Google's guidance above says. Give it a stable @id, which is just a fixed name other pages can point at, and reference that same @id from every other page on the site rather than describing your company again each time. The most common way this breaks on the About page specifically is two plugins each emitting their own organization block, so a machine sees two companies where there is one. Which Organization subtype you should actually be using, plain Organization, a LocalBusiness type, OnlineStore, or something else, is worth getting right and is covered in Organization vs LocalBusiness schema.

The people on the page can each be a Person entity, connected to the organization with worksFor and carrying sameAs links to their real professional profiles. Again, this is the same pattern as author entities, define each person once and reference them by @id.

The prose and the schema must agree. If the page says "founded in 2016" and the markup carries "foundingDate": "2018", you have handed a machine two answers and forced it to choose. Every field where the visible page and the structured data disagree is a small reason to trust the page less.

This is one page, not the whole job

Getting the About page right is necessary and it is not sufficient. It is the anchor that the rest of your identity signals point back to, but the work that really decides whether AI systems recognize you happens across the wider web: a Wikidata entry, a consistent sameAs network, genuine third-party references, demonstrated authority on specific topics. That full picture is laid out in the guide to your brand's entity profile. And if an AI assistant already describes your business wrongly or not at all, why ChatGPT doesn't know you is the diagnostic to run. This post is about the single page that all of that connects to.

Common mistakes

The anonymous About page. No names anywhere. "Our team of dedicated experts" and a stock photo. A business with no people on its About page has declined to answer the most human version of "who are you."

The origin story with no present tense. Three paragraphs about a kitchen table in 2011 and not one clear sentence about what the company does today.

A separate, disconnected Team page. The people exist, on their own page, which the About page never links and no schema connects to the organization. The data is real and it is doing no work.

Stale leadership. A founder who left two years ago still listed as CEO. A headcount from a funding round that has since doubled or halved. Facts that were true once and were never revisited.

Identity in the schema but not the prose, or the reverse. The markup has a clean foundingDate, founder, and legalName, and the visible page mentions none of them. Or the page states everything and the markup is a bare Organization with a name. Both surfaces need to carry the facts.

The About page that is actually a sales page. Testimonials, a pricing table, and a signup form, with no plain statement of who is behind it. That page answers "why should I buy" and leaves "who are you" open.

Facts trapped in an image, a PDF, or a script. The one that quietly costs the most, because the page looks complete to you and reads as empty to a crawler.

How to check your own About page

1. Read it once, cold, and write down the eight facts. Whatever you cannot answer from a single read is what a model cannot answer either.

2. Read only the headings. Do they map to questions a person would ask, or are they one-word labels? Rewrite the labels as questions.

3. Reload with JavaScript disabled, or just view the page source. Is the team still there? The founding date? The description? Anything that vanishes was only ever visible to browsers, not to crawlers.

4. Search the source for @type. Count how many blocks describe your company. More than one, with different @id values and no relationship between them, is a split identity to fix.

5. Line up the prose against the schema. Name, founding year, location, people. Every mismatch is a point against the page.

6. Ask an AI assistant directly. "Who runs example.com? What do they do, and where are they based?" If the answer is vague or wrong and your About page has the facts, the facts are on the page but not yet extractable, so work on the structure. If your About page does not have the facts, that is the place to start.

AI Schema Gen reads your actual site content and builds a connected entity profile from it, your organization, the people behind it, and the links between them, generated as one coherent graph rather than a pile of separate blocks. See what your site tells AI about you, free at aischemagen.com.