AI-SEO: how to get mentioned in ChatGPT answers

Three pillars determine whether ChatGPT mentions you: citable structure, demonstrable authority, and presence where AI systems look.

← back to home

AI visibility // approach
ai-seo.md

$ build --topic=ai-seo --pillars=structure,authority,presence

AI-SEO: how to get mentioned in ChatGPT answers

AI-SEO is the craft of getting mentioned in AI answers, and it requires three things: content with a structure that lends itself to citation, authority in the form of sources and figures, and presence in the places where AI systems look. Guarantees don't exist, but each of the three can demonstrably be improved and measured. This piece describes the approach; you can do most of it yourself.

DD DataDrift Digital • 14 September 2026 • 6 min

ChatGPT assembles its answers from texts it can read and cite. The question, then, is not how to put the model in your favour, but how to make material it can actually use. This work is now called AI-SEO, and it breaks down into three pillars, all three of which are simply work rather than magic. Get all three in order and there's nothing special left to do; skip one and you'll keep wondering why the competitor gets mentioned.

PillarWhat it requiresFirst action
StructurePassages that answer a question on their ownPutting the answer at the top of your most important pages
AuthorityFigures with a source, a named authorBacking up claims and putting a name under your pieces
PresenceAccurate mentions in the places where your market is describedChecking and updating existing mentions

The measurement tells you which pillar deserves priority for you: so start with a baseline measurement, not with building.

01 / structureMake your content citable

AI systems don't cite websites, they cite passages: standalone chunks of text that answer a question on their own. The first pillar is therefore that your pages consist of such passages.

AI doesn't cite websites. AI cites passages.

Two interventions do most of the work here. Put the answer at the top: the question in the heading, the full answer in the first paragraph, and only then the explanation. And give each piece a form that can be parsed: headings that match real search queries, a table where something is compared, a list where something goes step by step. The full writing rules are set out in how to write content AI can cite.

The good news for smaller players: this is the pillar where established names regularly drop the ball. Many large sites carry years of stacked marketing copy in which the answer is never at the top. As a result, one well-structured page addressing one real customer question can be chosen over a domain that's much stronger on paper.

02 / authorityGive the model a reason to choose you

Among ten pages making the same claim, an AI system chooses the page that backs up its claims. Princeton's GEO research (KDD 2024) measured this: citing sources, adding quotes and adding statistics together produced 30 to 40 percent more visibility in the generated answer, in a lab test; up to 37 percent on Perplexity.

Here too, two interventions are the starting point. Back your claims with figures and name the source with the year, even when the figure isn't your own. And put a name under your pieces: a named author with a recognisable background carries more weight than an anonymous company blog. What you leave out matters too: inflated claims without a source make a page less citable.

03 / presenceBe where the model looks

ChatGPT doesn't draw its picture of a market only from company sites, but mainly from the places where that market is described: trade media, review sites, comparisons, forums. Getting your own site in order is thus half the work; the other half is being present in the sources covering your field.

Concretely, that means: make sure mentions of your company on third-party sites are accurate and up to date, and genuinely take part where your field is discussed. Fake mentions and bought praise get found out and have no place here.

04 / the orderWhere to start

Start with measuring, not with building. Fix a set of search queries, note where you currently do and don't appear, and you'll know which of the three pillars is weakest for you. Without that baseline measurement, every improvement is invisible and every investment a gamble.

After that, the logical order is: first the structure of your most important pages, because that's work you do once and which stays in place. Then the substantiation, piece by piece. Presence with third parties runs through it as the slow work: it builds up over months and can't be forced.

Keep the pace realistic. Five pages that genuinely answer the most important customer questions do more than fifty written just to have a presence. And don't expect movement within a few weeks: indexing, recrawling and updating answers take months, and that's true for everyone, not just you.

05 / the honestyWhat nobody can promise you

Positions in AI answers cannot be guaranteed, by anyone. Answers differ per session and per platform, and the selection mechanism belongs to the model builder, not to you or an agency. What can be done: carrying out the method, recording the measurement and showing the movement in black and white. Why that's the only honest offer in this field is explained in can positions in AI answers be guaranteed. Anyone promising you otherwise is selling something they don't own.

Frequently asked questions
What determines whether ChatGPT mentions your business?+
Three things together: whether your content has a structure that lends itself to citation, whether your claims are backed up with sources and figures, and whether you're present on the third-party sites where your market is described. You can work on all three, and all three are measurable.
What does backing up content demonstrably deliver?+
Princeton's GEO research (KDD 2024) measured the effect of interventions on visibility in the generated answer: citing sources, adding quotes and adding statistics together produced 30 to 40 percent more (lab test), up to 37 percent on Perplexity. Substantiation is thus the best-documented intervention you have in your own hands.
Where do you start if you're not mentioned anywhere?+
With a baseline measurement: a fixed set of search queries, noting per query whether you appear and who does. Then tackle the structure of your most important pages first, then the substantiation, while bringing third-party mentions into order in parallel. Without a baseline measurement, every improvement stays invisible.
Does it help to pack lots of keywords into your text?+
No, it backfires. The same Princeton research (KDD 2024) found that keyword stuffing performed 10 percent worse on Perplexity. AI systems choose readable, substantiated passages, not text stuffed for an algorithm.
Can an agency guarantee a mention in ChatGPT?+
No. The selection mechanism belongs to the model builder and answers differ per session, so nobody can guarantee a position. What a serious party can guarantee: a fixed measurement method, a baseline measurement upfront and monthly re-measurements that show the movement in black and white.
baseline measurement · fixed price · no obligations

Want to know which of the three pillars is weakest for you?

The scan

This is the baseline measurement: fixed search queries on ChatGPT, Gemini, Claude, Perplexity and Google. Current prices are on the pricing page. Even without a follow-up, you keep the report.

→ The scan · pricing

This text was produced with AI assistance and checked and approved for publication by a human.