Consistent entity
The same business name, description, address or service area, phone and URL across the site, the profile, Companies House and your social profiles. Organization schema with sameAs links where they are real.
Guide
AI answers are influenced by the same things that make a business legible to any search engine: consistent facts, a real entity, structured data and content that answers the question. Most of the “AEO” tactics sold today are either unproven or actively misleading.
Answer engine optimisation, generative engine optimisation and AI search optimisation are three names for one question: how does a business get cited by an assistant? The honest answer is that the mechanisms are still developing, the measurement is poor, and the fundamentals are unglamorous.
An assistant needs to identify an entity, verify a fact and find a source. That means consistent naming, a real address and phone, a complete Google profile, a crawlable site with clear pages, and content that states the answer plainly. No amount of “llms.txt” fixes a business that cannot be identified.
What works
The same business name, description, address or service area, phone and URL across the site, the profile, Companies House and your social profiles. Organization schema with sameAs links where they are real.
The direct answer in the first 60 words, then the detail. This is good writing for people and the only format an extractor can reliably quote.
Service, areaServed, FAQPage and BreadcrumbList schema. Tables for prices and comparisons. Dates on guides. Facts in a form a machine can parse.
Local AI answers lean heavily on the profile, reviews, proximity and categories. There is no shortcut around it.
Local press, chambers, directories, client sites and industry bodies. The same links that help conventional search help an assistant decide you are real.
Track AI referrals as their own channel in analytics. Expect them to be small and noisy. Do not replace call tracking with an assumption.
What does not
Reality
| Channel | Typical role | What to measure |
|---|---|---|
| Google organic | The main source of commercial enquiries | Impressions, clicks, enquiries by page |
| Local pack | Calls and direction requests | Profile calls, directions, website clicks |
| AI Overviews | Informational queries, some commercial research | Referrals where identifiable; brand search lift |
| AI assistants | Research and recommendation | Referral traffic tagged by source; direct brand queries |
| Direct and referral | The compounding effect of being known | Brand search volume, direct sessions, repeat enquiries |
Practical
Fix the entity first: profile, NAP, schema, consistent description. Then make every important page answer its question in the first paragraph and carry a date. Add a visible FAQ with real questions. Publish one genuinely useful piece of local data or a guide, and earn a handful of local mentions. Measure referrals separately and do not panic when they are small.
That is not a new discipline. It is the same work as good local SEO, with a different extraction layer on top. The businesses that benefit will be the ones that were already legible.
Questions
It is a layer on top. The entity, the crawlable site, the structured facts and the citations all come from SEO. The difference is formatting for extraction: answer-first, dated, specific.
It is harmless and cheap. It is not a ranking tactic, and we would not charge for it as one.
Tag referrals by source in analytics, watch brand search volume, and ask new enquirers how they found you. The data is imperfect; the conversation is not.
For simple factual questions, sometimes. For a considered purchase — a builder, a solicitor, a hotel — people still check the site before they call. The site's job is to be the source that survives the check.
Tell us what you are trying to achieve. We will tell you honestly whether we can help, what it would take, and roughly what it would cost — before anyone signs anything.