AEO strategy: how to build one and run it in 90 days

What an AEO strategy decides

Answer engine optimization (AEO) is the work of getting a brand named and linked inside AI answers, which I explain in what is answer engine optimization. A strategy turns that work into a plan by answering four questions.

• Which questions should we be named in?
• Who gets named there today?
• Why them, and what is keeping us out?
• What do we do first, and how will we know it worked?

Google’s position is that optimizing for generative AI search is optimizing for the search experience, and thus still SEO [1]. The evidence backs the overlap and shows its limits. In Ahrefs’ analysis of 863,000 keyword results and 4 million AI Overview URLs, 37.9% of cited URLs also appeared in the first 10 results on the page, and 31% appeared beyond the first 100 [2]. So I run SEO and AEO as one program, and the strategy adds a second scoreboard: whether assistants name you, whether they link to you, and who appears instead.

Step 1: choose the prompts and the competitors

Everything later is measured against this list, so choose it first and keep it stable. Take the questions from sales calls, support tickets and Search Console queries, then add the prompts buyers would use at each stage of a purchase. Start with 20 to 50: enough to show patterns, few enough to read the answers yourself. Name the competitors you lose to, and add any the assistants keep naming.

Here is how the set might look for a project management tool built for architecture firms.

Stage	Example prompt	What you want to see
Category	Best project software for architects	Named in the list, with a link
Comparison	Your tool vs a rival for small firms	An accurate account of the difference
Problem	How do I track billable hours by project?	Your guide cited as a source
Brand	Is your tool good for a 10-person firm?	Facts that are correct and current

Step 2: measure a baseline before you change anything

Run the set across the assistants your buyers use, several times each. SparkToro’s volunteers ran 12 prompts 2,961 times and found under a 1 in 100 chance that ChatGPT or Google’s AI gives the same list of brands in any two responses [3]. So record how often you appear across runs, and treat a single answer as an anecdote.

For each prompt, record four things: named or missing, linked or unlinked, who appears instead, and which pages are cited. I track these across ChatGPT, Perplexity, Gemini and Google AI Overviews, because the share-of-voice figure they produce is what the strategy gets judged on.

Then add what the platforms report themselves. Search Console’s Generative AI performance report shows your impressions in Google’s AI features by page, device and country [4]. Bing Webmaster Tools shows how often your content is cited in Copilot and Bing’s AI summaries [5]. Google Analytics’ AI Assistant channel counts visits from ChatGPT, Gemini, Copilot and others, and leaves out AI Overviews and AI Mode [6]. I set out the full method in how to measure AI search visibility, and a quick check in how to tell whether AI is citing your brand.

Step 3: find out why someone else gets named

Read the baseline as a diagnosis before anyone writes a page. Each pattern has its own cause, and the cause sets the first move.

What you see	Likely cause	First move
Missing, even on brand prompts	Access or indexing	Check crawling, indexing and crawler rules
Described vaguely or wrongly	Entities	Align your facts across your site and profiles
Named on brand prompts, missing on category prompts	Topic coverage	Give each intent one strong page
Rivals cited from lists and reviews you are absent from	Mentions	Earn a place in those sources
Cited, with few visits or leads	The page or the tracking	Fix the page’s next step, then the source tracking

Most of these gaps are still open to close. In Semrush’s study of 1,094 ChatGPT categories, only 15.2% had a clear owner, meaning one brand in at least four of five prompts with a lead of five points [7]. For a leading news portal, share of voice in AI answers went from 9% to 41% against its top 5 competitors, with 120+ queries moved into cited positions through entity coverage. I cover that work in entity SEO for AI search.

Step 4: put the work in order

Work through the layers in this order, because each one depends on the one before it. Skip any layer your diagnosis shows is already sound.

1. Access. A page must be indexed and eligible to show with a snippet to appear in Google’s AI features [1]. OpenAI and Perplexity search with their own crawlers, OAI-SearchBot and PerplexityBot, so check that your rules allow both [8][9].
2. Entities. Use one name, one description and the same facts on your site and your profiles. Structured data supports this, and Google says no special schema is needed for its AI features [1].
3. Content. Group the questions and the searches behind them by intent, and give each intent one strong page. Add what only you know: Google says this will likely influence your presence in generative AI search more than any of its other suggestions [1].
4. Mentions. Earn your place in the sources assistants already cite, through digital PR and real coverage. In Ahrefs’ study of 75,000 brands, YouTube mentions correlated with AI visibility at about 0.74 and branded web mentions at about 0.66 to 0.71 [10]. Those are correlations only, and Google warns that inauthentic mentions help less than they seem to [1].
5. Measurement. Keep every source apart in your conversion tracking, so you can see what the citations bring in.

Content and tracking pay off together. For one e-commerce brand, citations of its buying guides in AI answers drove +38% AI-referred sessions and +19% assisted conversions in 60 days. The planning side is in AI content strategy, and the searches behind a prompt are in how query fan-out works.

Step 5: review it every month

Re-run the same prompts on the same schedule and compare the trend with your baseline. Read each change as a decision: a prompt you won shows what worked, a prompt a rival took sets the next task, and a citation that brings no visits sends you back to the page.

Lead the report with the citation numbers, then the visits and conversions they brought in. Review the prompt set itself once a quarter, so it keeps matching what buyers ask.

A 90-day AEO plan

Here is the plan on a calendar. Ninety days gets you the fixes, the content plan and a first reading against the baseline.

When	The work	What you have at the end
Days 1 to 30	Prompts, baseline, competitors, technical audit, entity fixes	A ranked list of gaps, each with an owner
Days 31 to 60	Keyword research, content plan, briefs, first pages merged or rewritten	One page planned for each intent
Days 61 to 90	Publishing, first earned mentions, a second measurement	A second reading against the baseline
Every month after	New and refreshed pages, mentions, a report, prompts reviewed each quarter	A trend you can act on

This is the shape of the programs I run: a strategy and execution plan, then a content strategy and plan, then ongoing optimization, on the tools I build. The answer engine optimization page shows how that works for clients, and if you would rather hand it over, I set out what an AEO consultant does and how to hire one.

What to leave out of the plan

Google lists several tactics you can ignore for its Search [1]. Spend that time on the five steps instead.

• New machine-readable files or special markup, which Google Search does not use.
• Chunking content into tiny pieces. Google says its systems understand several topics on one page.
• Rewriting pages in a special style for AI. Write clear sections for your readers.
• A separate page for every variation of a search, which breaks Google’s scaled content abuse policy when done primarily to manipulate results.
• Inauthentic mentions, which its core ranking and spam systems work against.

Be careful with tools too. Google says no third-party tool has access to its internal ranking or AI systems [1], so judge a tool’s advice against what you can measure yourself.

A worked example: a project tool for architects

The company and its numbers here are illustrative. A project management tool for architecture firms tracks 30 prompts across three assistants, with five runs each. It appears in 3 of the 30. A larger rival appears in 14.

The diagnosis is specific. The answers call the tool a construction app, which is an entity problem. The rival is cited from three review sites where the tool has no profile, which is a mentions problem. Six thin posts compete on billable hours, which is a coverage problem.

In the first month, the team corrects its name and description across its site and profiles. In the second, it merges the six posts into one guide and adds what only it knows: how long projects really run, from 400 firms’ timelines. In the third, it earns a place on two of the review sites. At day 90, it appears in 8 of the 30, and the prompts it still misses set the next quarter’s work.

Frequently asked questions

Sources

1. Google Search Central, “Google’s guide to optimizing for generative AI features on Google Search”, last updated 10 July 2026. developers.google.com
2. Ahrefs, Louise Linehan, “Update: 38% of AI Overview Citations Pull From The Top 10”, 2 March 2026. ahrefs.com
3. SparkToro, Rand Fishkin, “NEW Research: AIs are highly inconsistent when recommending brands or products; marketers should take care when tracking AI visibility”, 28 January 2026. sparktoro.com
4. Search Console Help, “Generative AI performance report (Search)”, read 30 October 2026. support.google.com
5. Microsoft Bing Webmaster Blog, “Introducing AI Performance in Bing Webmaster Tools Public Preview”, 10 February 2026. blogs.bing.com
6. Google Analytics Help, “Default channel group”, read 30 October 2026. support.google.com
7. Semrush, Margarita Loktionova with Kevin Indig, “AI visibility is a topic-level game: A study of 50,000 brands in ChatGPT”, 20 July 2026. semrush.com
8. OpenAI, “Overview of OpenAI Crawlers”, read 30 October 2026. developers.openai.com
9. Perplexity, “Perplexity Crawlers”, read 30 October 2026. docs.perplexity.ai
10. Ahrefs, Louise Linehan, “Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied)”, 12 December 2025. ahrefs.com

Related insights