For twenty years, getting found online meant ranking on page one of Google. That still matters, but a growing share of people now ask an AI assistant instead. They type a full question into ChatGPT, Gemini, Perplexity or Google's AI Mode, and get one synthesised answer with a handful of cited sources.
If your business is one of those sources, you win the click and the trust. If it is not, you may never know the customer was looking. Generative engine optimisation (GEO), sometimes called answer engine optimisation (AEO), is the practice of making your website easy for these systems to find, understand and quote.
How AI assistants pick their sources
Most AI search experiences work in a similar way. The assistant rewrites the user's question into several search queries, retrieves a set of pages from a web index, reads the most relevant passages, and writes an answer that cites the pages it relied on. That process has a few important consequences:
- You still need to be indexed and crawlable. If search engines and AI crawlers cannot fetch your pages, you are invisible.
- Passages matter more than pages. The model lifts specific paragraphs, so each section needs to make sense on its own.
- Clear, factual statements get quoted. Vague marketing copy gives the model nothing to cite. Specific facts, definitions, prices, steps and comparisons do.
- Trust signals carry over. Authorship, reviews, consistent business details and mentions on other reputable sites all influence which sources the system prefers.
Write for questions, not keywords
Traditional SEO trained us to target short phrases like "recruitment software". AI users ask things like "what software do small labour-hire agencies use to manage timesheets and compliance?" The best way to show up is to answer those questions directly.
Start by collecting the real questions your customers ask. Look at sales call notes, support tickets, live chat logs and the "People also ask" boxes in Google. Then make sure each important question has a clear answer on your site, ideally in the first one or two sentences under a heading that matches the question.
Structure content so machines can parse it
Large language models are good at reading messy text, but the retrieval systems that feed them still reward clean structure. A few habits make a real difference:
- Use one
h1per page and a logicalh2/h3hierarchy that reads like a table of contents. - Prefer HTML text over text baked into images, sliders or PDFs.
- Use real tables for comparisons and real lists for steps.
- Keep key content in the initial HTML response. Content that only appears after heavy JavaScript runs may never be seen by some crawlers.
- Add a short summary or key takeaways at the top of long articles.
Add structured data (schema)
Schema.org markup tells machines exactly what a page is about: an organisation, a service, a product, an article, an FAQ or a how-to. It removes ambiguity about your business name, location, contact details and offerings. At a minimum, most business sites should have:
| Schema type | Where to use it | Why it helps |
|---|---|---|
| Organization / LocalBusiness | Home page or site-wide | Confirms your name, logo, contact details and social profiles |
| Service | Each service page | Describes what you offer and where |
| Article / BlogPosting | Blog posts | Signals authorship and publication date |
| FAQPage | Pages with genuine Q&A | Packages question and answer pairs cleanly |
| BreadcrumbList | All inner pages | Explains how pages relate to each other |
Use JSON-LD, keep it accurate, and make sure everything in the schema is also visible on the page.
Technical checks for AI crawlers
AI companies run their own crawlers alongside traditional search bots. Review your robots.txt and any firewall or bot-protection rules to make sure you are not blocking the crawlers you want, such as those used for search and citation, by accident. Decide deliberately which bots you allow for training versus search, because many providers now separate the two.
Some sites also publish an llms.txt file: a plain-text summary of the site with links to the most important pages. Support for it varies, but it is cheap to add and forces you to articulate what your site is about. Fast load times, a clean XML sitemap and correct canonical tags remain essential.
Build authority beyond your own website
AI assistants cross-check information. If your business is described consistently on your website, your Google Business Profile, industry directories, review platforms, LinkedIn and in genuine press or partner mentions, the model is more confident recommending you. Practical steps include:
- Keep your name, address, phone and service descriptions identical everywhere.
- Earn reviews on the platforms your customers trust.
- Publish original insight, such as data, case studies or expert opinion, that others want to reference.
- Show real authors with real credentials on your content.
How to measure GEO
Measurement is still maturing, but you are not flying blind. Check referral traffic from AI platforms in your analytics, watch for branded search growth, and regularly ask the major assistants the questions your customers ask, then record whether and how you are mentioned. Dedicated AI visibility tracking tools can automate this across many prompts and models.
GEO is not a replacement for good SEO. It is good SEO, written for a reader that happens to be a machine summarising for a human.
A quick GEO checklist
- List the 20 questions customers ask most, and answer each one clearly on your site.
- Use answer-first paragraphs under question-style headings.
- Add Organization, Service, Article and FAQ schema where relevant.
- Make sure AI search crawlers can reach your pages.
- Keep business details consistent across the web.
- Track AI referrals and test your visibility monthly.
If you would like a GEO audit of your website, including schema, content structure and AI visibility testing, get in touch. It is part of our AI strategy and integration work.
