AEO Maturity Model: Five Stages and What to Fund

We checked the crawler, structured data and Search Console details below against Google's, OpenAI's and Anthropic's own documentation on 14 October 2026.

What is an AEO maturity model for?

An AEO maturity model tells you which problem to pay for next, so budget goes to the gap holding you back.

Each stage is defined by what you can see in ChatGPT, Claude and Gemini answers: whether you are named, for which personas and topics, whether you are recommended or just listed, and whose pages get cited. Rerun it each quarter.

What are the five stages of AEO maturity?

The five stages are Invisible, Aware, Active, Established and Industry Default, and each calls for a different investment.

The share bands are an illustrative mapping to the competitive tiers in our citation share benchmarks, measured as mention-based share of voice.

The "Not yet" column is our editorial view of sequencing, not a sourced finding. When you straddle two stages, fund the lower stage's open gaps first, because later-stage work pays off only once engines can read, quote and identify you.

How do you tell which AEO maturity stage you are at?

The stage comes from the pattern across many answers; one screenshot proves nothing.

1. Fix the question set. Write the buyer questions once, per persona, and keep them stable between runs. We research each persona from the brand's market because each persona gets a different answer.
2. Run each engine separately. Our AI visibility platform runs ChatGPT, Claude and Gemini by default with web search, can add DeepSeek and Grok, and runs every prompt twice per engine.
3. Separate three outcomes. Count answers that recommend the brand apart from answers that only mention it or only link to it.
4. Note whose pages are cited. Your own, a third-party listing or a competitor's.
5. Hold the scoring method constant. Each monthly report we send tracks LLM recall frequency, mention rate and citation share. Change the method midway and the trend line stops meaning anything.

Score each slice on its own, since a brand can be Established for its flagship product at home and Invisible for a newer product line, a second country or a second persona. The full setup is in how to measure your AI visibility.

How do you know you are at Stage 1, Invisible, and what should you fund?

You are at Stage 1 when AI engines answer your category questions without naming you. Fund the access and pages that let engines read you, plus a record of today's answers.

Signs you are here

1. Do you appear only when the question already includes your brand name?
2. When you are cited, is the link a directory or listing, never your own site?
3. Does your robots.txt, CDN or firewall block AI search crawlers?

Mostly yes means Stage 1.

What to fund next

• Crawler access. Let crawlers through at every layer: robots.txt, the CDN and the hosting setup, as Google's AI features guide asks. According to OpenAI's bots page, ChatGPT's search features use OAI-SearchBot to show sites in their results, you control it apart from OpenAI's other crawler settings, and a site that opts out is left out of ChatGPT search answers except as a navigational link. Anthropic's crawler page says disabling Claude-SearchBot may reduce your visibility in Claude's search results. The exact rules are in our robots.txt guide.
• Indexable answer pages. For a supporting link in AI Overviews or AI Mode, Google's bar is a page that is indexed and eligible to show in Search with a snippet. Beyond that, the guide says: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."
• A baseline. Save today's answers and the brands they name before you change anything.

How do you know you are at Stage 2, Aware, and what should you fund?

Engines name you sometimes, for some questions, but a buyer cannot count on seeing you. Fund one clear answer page for each buyer question you want to win.

Signs you are here

1. Are you named on one engine but missing on the others for the same question?
2. Does one persona see you while another, asking a close variant, does not?
3. When you are named, are you one entry in a list with no recommendation?
4. Do third-party listings get cited about you more often than your own pages?
5. Did your first few mentions vanish when you reran the full question set?

What to fund next

• One page per buyer question. Answer it in the opening lines. Google's helpful content guidance lists extensive automation to produce content on many topics as a warning sign, so buy coverage of real questions and skip bulk output.
• Markup that matches the page. Google's AI features guide says no special schema.org structured data is needed for AI features. Google's structured data introduction says not to mark up information users cannot see, and recommends JSON-LD where your site allows it.
• Search Console tracking. Traffic from AI Overviews and AI Mode is counted in the Performance report under the Web search type, so watch the pages you build there.

How do you know you are at Stage 3, Active, and what should you fund?

You are at Stage 3 when engines name you reliably for your core questions. Fund coverage for every persona and one consistent description of the company.

Signs you are here

1. Can you predict which core questions will name you, and be right most of the time?
2. Are you strong for your main persona but thin for the others?
3. Do engines describe what you do inconsistently, or mix you up with a similar name?

What to fund next

• Persona coverage. Build answer pages for the personas where you thin out, and break citation share down by persona to see which ones those are.
• One clear entity. Use the same name, category and one-line description on your site, profiles and listings. The steps are in building a brand entity AI engines recognize.
• Accurate bylines. Google strongly encourages accurate bylines where readers expect them, and calls fabricated creator profiles a form of deception.
• First earned mentions. Start with the third-party pages engines already cite where you are missing; our platform ranks those sites for each client.

Hold category-defining research and executive programs until engines know who you are, or the coverage has no clear entity to attach to.

How do you know you are at Stage 4, Established, and what should you fund?

Stage 4 means ChatGPT, Claude and Gemini recommend you for most of your personas. The bottleneck has moved off your site, so fund mentions on the sites engines already trust.

Signs you are here

1. In "best X for Y" answers, are you recommended as well as listed?
2. Are your remaining gaps questions where engines cite review sites, listings or forums that never mention you?

What to fund next

• Earned mentions. Our team earns mentions on the reviews, listings, forums and publications AI already reads in a client's category, through monthly PR, publishing on other sites and community seeding.
• Comparison answers. Answer "X vs Y" and "alternatives to X" honestly on your own site, or engines take the framing from someone else.
• Upkeep of cited pages. We re-score existing articles and queue the weak ones for rewrite.

A New York edibles retailer reached an Established-like position inside a single program cycle: from 1.5% AI visibility at #50 to 25%, #4 of 374 brands. By the end, the retailer's site was cited more often than any other source in the category, the state government's official site included.

How do you know you are at Stage 5, Industry Default, and what should you fund?

At Stage 5, engines give your brand as the default answer to category questions. Defend the answers you have won, then fund the next product line or market.

Signs you are here

1. Asked for the best option in your category, does each engine you track name you first?
2. Do engines explain the category in your terms?
3. Is your site cited as the source for category definitions?

What to fund next

• Defense. Rerun the question set on schedule and treat every lost answer as work for the next cycle.
• Updates with substance. Update cited pages when the facts change. Google lists changing page dates without substantial changes as a warning sign.
• Expansion. Apply the model again to adjacent products, markets and personas.

In our view, Stage 5 reflects market position as much as content. AEO can make a category leader visible; it cannot create the leadership.

What AEO investment mistakes should you avoid at each stage?

The costly mistakes are timing mistakes: funding a stage's work too early or too late.

• Too early. Paying for PR, entity projects or executive thought leadership before the Stage 1 basics are in place, when our benchmarks post says the fastest early gains come from technical discovery and structure, not volume.
• Too late. Paying for more format work on pages engines already cite after the gap has moved to other sites (see Stage 4).

How long does each stage take?

No universal clock exists, so this model gives no month counts. OnlyAEO's guidance is that first measurable results usually need at least 60 days; after that, visibility compounds over the following months as engines pick up more pages, videos and mentions.

Google's structured data introduction suggests recording performance for a few months in the Performance report before judging markup.

Our benchmarks post also warns against judging a content push before its second full re-index. Later stages are measured in quarters, because they depend on other sites deciding to mention you. Engine-by-engine detail is in how long AEO takes, engine by engine.

How we tested this

On 14 October 2026, before writing, we asked ChatGPT and Claude this exact question and read the pages each one cited. Claude's answer cited four maturity models, from AEO vendors, a website platform and a press release.

Those models define stages by tactics or by mention-rate bands that disagree with each other, and none ties a stage to what to fund next. We built each stage here around that question.

Where does OnlyAEO fit?

OnlyAEO is an AEO and GEO agency. The managed service bundles monthly AI visibility analytics, YouTube videos, articles in other languages, researched and fact-checked articles in your voice and design, help getting your brand onto trusted sites and listings AI reads, and a website health check.

As how it works lays out, each month a fresh report shows where you are named, we create work you approve, and we review together what worked and plan the next month. We never promise particular rankings, citations or placements in AI answers, on that timeline or any other.

More questions about AEO maturity

Can a brand slip back a stage?

Yes. Our benchmarks post says citation share drifts 40 to 60% month over month as engines re-index and re-rank, so one good month does not settle a stage.

Do we need an llms.txt file to move up a stage?

Not for Google. Google's AI features guide lists AI text files among the things you do not have to create for AI Overviews or AI Mode, along with new machine-readable files and markup. Treat llms.txt as optional.

Should we judge our stage on Google AI Overviews alone?

No. Google says AI Overviews appear only when its systems judge them additive to classic Search, so they often don't trigger for a given question.

When can we stop funding the next stage?

Our benchmarks post says to deprioritize further citation-share work when you are consistently top three on citation share for your highest-intent prompts, your AI share matches your market share, and the marginal cost exceeds the pipeline return.