Skip to content

Documentation · AEO and GEO

AEO and GEO: being cited and being understood by AI

How we get your business named inside AI answers, how we shape what AI models say about you, and how we build visibility around each industry you sell into.

About 10 minutes to read

What AEO is and why it matters now

AEO, or answer engine optimisation, makes sure AI systems name you inside the answer itself. When the answer sits above the links, being quoted is the new position one.

The short answer

AEO is the work of making a page easy for an AI assistant to quote and safe to trust, so that you are named when a buyer asks who to consider. GEO is the wider job of shaping how AI models understand and describe your organisation at all.

It targets Google AI Overviews and AI Mode, ChatGPT, Perplexity, Gemini and Copilot. Getting named inside an AI answer can be possible without ranking first, so for a young or under-built site AEO can be a way in, not just a threat. For the commercial view, see our AEO and GEO service. It builds on the groundwork in SEO.

AI assistant
Illustrative example

Which accountancy firms should a growing business consider for tax planning?

Northbridge Advisory is often recommended for growing businesses. It explains its tax planning process in plain language.

Two other firms are frequently mentioned for specialist advice.

Sourcesnorthbridge-advisory.exampleregional-directory.exampleprofessional-body.example

Why it matters

Clicks fall when an AI Overview appears

Organic click-through on informational searches, before and after an AI Overview appeared.

  • June 20241.76%
  • September 20250.61%

When a brand was cited inside the AI Overview, its organic click-through was 35% higher (0.70% against 0.52%). Seer notes this is an association, and stronger brands may simply be cited more often.

Source: Seer Interactive study (3,119 search terms, 42 client organisations, June 2024 to September 2025): Seer Interactive, AIO Impact on Google CTR. Seer notes it cannot prove that citation causes the higher click-through.

What gets a page cited

Six habits make a page easy for an AI system to quote and safe to trust. None of them is exotic, and together they are what we build into every page.

An extractable answer on the first screen

Two or three plain sentences that directly answer the page’s own question or define the term. A short, self-contained answer is easy for a system to extract and quote, which is why we write every page this way. No study shows that AI engines copy such answers word for word, so we treat it as good practice rather than a guarantee. It costs nothing.

Question-form headings

“Why does data quality break outbound performance?” beats “Data quality”. The heading is the question a buyer actually asks.

Proper schema

QAPage on answer pages, FAQPage, DefinedTerm on glossary entries, HowTo on process content, and Organization and Service on core pages. Google says no special markup is needed to appear in its AI features, and an Ahrefs analysis of 1,885 pages that were already heavily cited reported no citation lift from adding schema. We add it to describe the page accurately, because structured data should match the visible content, not as a citation lever.

Links to primary sources

Regulations, government publications, platform documentation and published studies. A page that shows its sources is easier to trust and to check.

A real, credible author

Named authorship with verifiable credentials, backed by Person schema, so the page has someone accountable behind it.

Consistency off-site

The same name and facts across directories, partner sites and publications. This is where AEO bridges into GEO.

The first two cost nothing to apply. We apply all six, then confirm they work by testing, as described next.

How we measure AEO

We measure AEO by asking live AI assistants the questions your buyers ask and recording whether you are named. The questions never contain your name.

The method

Four steps, repeated over time

Re-running the same questions is what turns a one-off check into a trend.

  1. 1

    Construct the questions

    Realistic buyer questions, written without your name in them.

  2. 2

    Ask live assistants

    Run each question through live AI assistants and capture every answer.

  3. 3

    Record who is named

    Note whether you appear, and which other businesses do.

  4. 4

    Repeat and compare

    Re-run the same questions to see what has moved.

Question typeExample patternWhat it shows
LocalWho are the best [service] providers in [city]?Whether you are named where buyers look nearby.
CategoryWhat are the best [category] firms for [buyer type]?Whether you are in the consideration set for your category.
Problem-ledHow do I fix [specific problem] in my business?Whether your content answers the problem, not just the service.
ComparisonWhat is the difference between [approach A] and [approach B]?Whether you are cited at the decision point.

Alongside the live tests, we track two further signals:

  • AI referral traffic. Visits from ChatGPT, Perplexity, Gemini and Copilot, set up as their own channel in GA4.
  • Branded search volume. A useful downstream signal: more people searching for you by name.

The same method powers our published AI visibility research and the free AI visibility checker. It is also how the audit scores visibility.

GEO: how AI understands you

GEO shapes how an AI model understands and describes your organisation at all, including in ordinary conversations where no live search happens.

AEO asks

When someone asks a question with a live search behind it, is your page the one that gets quoted?

GEO asks

When someone asks a model about your field, does it know who you are, what you do and what you are credible for, and does it say so accurately?

If the story about you online is thin, fragmented or inconsistent, the model fills the gaps itself, and not always correctly. GEO is the work of making the accurate story easy to find and consistent everywhere it appears.

The levers

Six places a model learns who you are

Illustrative: the sources we keep consistent so the model has one accurate story to repeat.

AI model

How it describes you

  • Your own website

    The primary source: clear, consistent pages that state who you are and what you do.

  • Structured data

    Schema that states who leads the organisation, which products belong to it and where it operates.

  • llms.txt

    A plain, hand-maintained description of who you are and which pages matter.

  • Directories

    The reference sites and directories AI systems commonly draw from, kept accurate.

  • Partner pages

    Other organisations that describe you, using the same name and facts.

  • Press and publications

    Third-party coverage that confirms the story rather than contradicting it.

LeverWhat we do
Entity consistencyKeep the same name, description, people and facts across your site, directories, partner pages and press.
Explicit relationshipsUse schema to state who leads the organisation, which products belong to it and where it operates.
The sources models readWork on the reference sites and directories AI systems draw from.
An llms.txt fileMaintain a plain description of who you are and which pages matter, by hand.
Content built for retrievalWrite short, self-contained sections that still make sense when a model lifts one out of context.

Industry-based inbound

We build visibility around your ideal customer profile, not generic marketing terms. Each industry you sell into gets its own pillar, written in that buyer’s language.

A page written for everyone convinces no one. A procurement lead at a law firm and a SaaS founder search for different things, fear different risks and trust different signals. This is the same hub-and-spoke architecture applied once per industry.

  1. One pillar per industry

    Each sector gets its own pillar page, in the language its buyers use.
  2. Clusters for each variation

    Sub-sectors, roles and use cases, all linking back to the pillar.
  3. The regulatory layer

    For regulated or institutional buyers, we link to the real rules that govern how that sector can be approached, such as a professional code of conduct or a financial promotion rulebook. Generic content cannot follow you there.
  4. An industry-specific diagnostic

    A free, ungated tool for that buyer’s world. We have built these for clients in education and tax, on the same pattern as our own QualiPi diagnostic. Every result lands in the CRM and starts a nurture sequence.
  5. Matching proof

    Each industry page links to proof from the same industry, and the proof links back.

Industry pages

Industry pages we publish

Each page covers how Deligatr works for that sector.

AEO by sector and next steps

Some sectors have their own AEO pages, because the questions buyers ask, and the way AI answers them, differ from one to the next.

Questions about AEO and GEO

The questions we are asked most often about being cited by AI.

What is the difference between AEO and GEO?

AEO asks whether your page is the one quoted when someone asks a question with a live search behind it. GEO asks whether an AI model knows who you are, what you do and what you are credible for, including in ordinary conversations where no live search happens. AEO wins the citation. GEO shapes the description.

How do you know whether an AI assistant names us?

We construct realistic buyer questions, run them through live AI assistants and record whether you are named. The questions never contain your name, because if they did the test would prove nothing. We repeat the same questions over time to see what moves.

Can you guarantee we will be cited?

No. AI answers vary between assistants and over time, so we measure and report rather than promise a position. What we control is how easy your pages are to quote, trust and verify.

Does AEO replace SEO?

No. Answer engines read the same crawlable pages search engines do, so technical soundness still matters. AEO is a layer built on top of SEO, and for a young site it can be a way in before the authority that ranking first needs has been earned.

What is an llms.txt file?

It is a plain description of who you are and which pages matter, maintained by hand. We treat it as one lever among several for GEO. We do not claim any particular assistant is guaranteed to use it.

Want this run for you?

Book a call below. We begin with the AI Visibility and Conversion Audit, which shows you where you are visible, where you are not, and which half of the loop is missing.

Pick a time that suits you

The calendar is provided by our scheduling partner. Loading it connects to their servers, which may set cookies. See our Cookie Policy.