Optimising for AI agents, often called Answer Engine Optimization (AEO), is the practice of making a business discoverable, understandable and recommendable by AI systems. Where SEO optimised for a human clicking a result, AEO optimises for a model extracting, trusting and citing your information.
AEO vs SEO
SEO targets a human scanning results. AEO targets a model parsing content and choosing an answer.
SEO wants the click. AEO wants the citation, to be the source an agent quotes and the option it recommends.
Bad SEO means page two. Bad AEO means you are not in the answer at all.
Fast, accessible, well-structured, honest pages help both. AEO extends SEO; it does not discard it.
How to optimise for AI retrieval
- Answer first: put a direct, self-contained answer immediately after each question-style heading, so a passage makes sense when retrieved on its own.
- Structure everything: Schema.org for products, FAQs and organisations turns prose into facts a model can trust.
- Define your entities: state clearly what B2A, your brand and your products are, and link related concepts descriptively.
- Server-render key content: agents often do not run JavaScript; keep important information in static HTML.
- Publish an llms.txt: Jeremy Howard of Answer.AI proposed llms.txt in September 2024, a Markdown index of your key content; it is now used by Anthropic, Cloudflare and Vercel. Fact
- Allow the crawlers: permit GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended in robots.txt. You cannot be cited if you cannot be read.
What not to do
Do not keyword-stuff or attempt to manipulate model outputs. The objective is to become a trusted source worth retrieving, not to game it. Answer engines increasingly reward verifiable, well-sourced information. Analysis