What AEO means
Search engines return a page of links and let you choose. Answer engines, the AI assistants built into ChatGPT, Perplexity, Google, Gemini, Claude and Copilot, read the web for you and return an answer. When the question is "who should I hire," that answer is a short list of two or three businesses, each with a reason attached.
AEO is the work of getting onto that list, and of making sure the reason attached to your name is accurate. It's related to SEO, but the goal is different: not a position on a results page, but a mention inside an answer that has no results page at all.
Not named means not considered. There is no page two in an AI answer.
Why it matters now
The way people look for a lawyer, a surgeon or a contractor has changed quickly. A few markers of how far it has moved:
- 800 million+ people use ChatGPT every week, according to OpenAI (October 2025).
- Over 2 billion people a month see Google's AI Overviews, the AI answer above Google's search results, according to Alphabet (July 2025).
- Gartner forecast a 25% drop in traditional search volume by 2026 as people shift to AI assistants (February 2024).
The engines are still deciding who the trusted names are in each category and each city. Consensus is sticky: once several sources agree a competitor is the pick, every answer repeats it, and it gets harder to displace. Businesses that do this work early become the default answer.
How AI engines choose who to recommend
Each engine works a little differently, but the pattern is the same. When someone asks for a recommendation, the engine:
- Interprets the question: the service, the location, and what "best" means to this person (price, results, speed, bedside manner).
- Retrieves sources from its own index or a live search: directories, review sites, news, forums, roundups and business websites.
- Looks for consensus: which businesses appear repeatedly, with consistent facts and positive, specific descriptions.
- Writes the answer, naming the businesses it's most confident about and explaining why, often in phrases lifted from reviews and directory profiles.
Two details matter. First, the engines don't share an index: ChatGPT search, Perplexity and Claude use their own crawlers, while Google's AI Overviews draw on Google's index and Copilot on Bing's. Ranking well in one says little about the others. Second, AI is confident about what's consistent. A firm with the same name, address, practice areas and strengths everywhere gets recommended over one whose profiles disagree.
AEO vs SEO vs GEO
| SEO | AEO / GEO | |
|---|---|---|
| Goal | Rank a page for a keyword | Be named and described inside an AI answer |
| Where it shows | A results page with ten links | A single answer with two or three names |
| What decides it | Your pages, links and technical health | Consensus across your site, directories, reviews and mentions |
| How it's measured | Rankings, clicks, traffic | Share of tracked questions where each engine names you, and how it describes you |
| Time to results | Months | First movement in weeks, stable in about 90 days |
GEO (generative engine optimization) comes from academic research into how content is used by generative AI. AEO is the more common industry term. You'll also see "LLM SEO," "AI SEO" and "AI search optimization." They describe the same goal.
The six signals that decide who gets named
1. Engines can read you
AI crawlers such as OAI-SearchBot (ChatGPT search), PerplexityBot and ClaudeBot each follow their own rules in your robots.txt file. Old rules often block one by accident. Sites that only show content after JavaScript runs can hand crawlers an empty page. Fixing access is the first step.
2. Answer-shaped pages
Pages that state the facts a model needs in plain sentences it can quote: what you do, for whom, where, how much, how long, and why you're the right choice. FAQ sections and clear service or practice-area pages do most of the work.
3. One consistent identity
Your name, address, phone, services, people and credentials matching across your website, Google Business Profile, directories and structured data (schema). Conflicting facts make the engines hedge.
4. Review consensus
Not just a high star rating, but recent reviews that repeat the same specific strengths: "returned my call within an hour," "natural results," "showed up the same day." AI quotes the pattern.
5. Third-party mentions
Best-of roundups, industry directories, local news, professional associations, Reddit threads and podcasts. These are the sources engines lean on to decide who's credible.
6. Measured, not guessed
A fixed set of real customer questions, run on every engine on a schedule, with the raw answers kept. Without it, AEO is managed on whichever screenshot someone took last.
AEO by industry
Law firms. People ask "best personal injury lawyer near me" or "do I need a lawyer for a minor accident." The engines lean on Avvo, Justia, Super Lawyers and Martindale-Hubbell, Google reviews that mention case types and communication, and clear practice-area pages. All of it has to respect attorney advertising rules. AI visibility for law firms →
Medical and aesthetic practices. Patients ask "who's the best surgeon for a natural-looking rhinoplasty" or "where should I get Botox." Engines lean on Healthgrades, RealSelf, Zocdoc and provider credentials, and they hold health answers to a higher bar. No patient data is ever needed. AI visibility for medical practices →
Home services. Homeowners ask "most reliable HVAC company for a same-day repair." Engines weigh service areas, licensing, response times and reviews across Google, Angi, BBB and Nextdoor.
How to measure AEO
Measurement turns AEO from guesswork into a channel you can manage:
- Build a question set. 25 to 150+ real questions customers ask, grouped by service, practice area or procedure and by market.
- Run it on every engine. ChatGPT, Perplexity, Google AI Overviews, Gemini and Claude, on a fixed schedule, from the right locations.
- Record the raw answers. Whether you're named, your position, how you're described, and which sources were cited.
- Score it. Your AI visibility score is the share of tracked questions where at least one engine names you. Track it by engine, market and competitor over time.
Getting started
The fastest first step is a baseline. Ask the engines the five questions your best customers ask, note who's named, and look at which sources the engines cite. If you're not there, those sources are your to-do list.
Or let us do it: our free AI visibility audit runs your real questions live across the engines in 20 minutes and shows you who's named, who isn't, and why. You keep the baseline whether or not you work with us.