Step 01
Baseline
Ask the assistants your customers questions and record where you stand.
Get cited by ChatGPT, Perplexity and AI Overviews.
Answer Engine and Generative Engine Optimisation - structuring a site so AI assistants can read it, trust it and cite it when they answer questions in your category.
A growing share of searches now end in an answer rather than a click. If an assistant answers the question your page was ranking for, the traffic never arrives - unless your site is the source the answer is built from.
That requires different work from classic SEO: machine-readable structure, answer-first writing, explicit crawler permissions, and a plain-text summary of the business that a model can ingest without parsing rendered markup.
Classic SEO optimises for a ranked list of links. Answer engines synthesise a response from sources they can parse and trust. A page can rank well and still never be cited, because its content is structured for human scanning rather than machine extraction.
The practical difference is concentration: an assistant lifts a passage. So each section has to stand alone, lead with the direct answer, and carry enough context to be quoted without the surrounding page.
GPTBot, ClaudeBot, PerplexityBot, Google-Extended and others each respect robots.txt separately. Many sites block them by accident through a blanket rule, then wonder why they are never referenced.
If being cited is the goal, those crawlers need explicit permission. I set this up deliberately and document it, so the choice is recorded rather than inherited from a template.
A plain-text summary of who you are, what you offer, your pricing and your common questions lets a model answer accurately without reconstructing it from markup. It is cheap to produce and removes the most common source of wrong AI answers about a business.
Paired with a validated schema graph, it gives assistants two independent, agreeing descriptions of the business - which is exactly what raises confidence enough to be cited.
There is no position one in an answer engine. Measurement means sampling: asking the assistants the questions your customers ask, recording whether you appear, and tracking that over time against competitors.
I set up that baseline so the work is accountable. It is less precise than rank tracking and considerably more honest than claiming an unmeasurable result.
Answer Engine Optimisation and Generative Engine Optimisation both mean structuring a site so AI assistants can parse, trust and cite it. AEO leans toward direct question answering, GEO toward being a source in generated responses. In practice the work overlaps heavily.
It builds on the same foundation but optimises for a different outcome. Classic SEO wins a position in a list; AI optimisation wins a citation inside an answer. Valid structured data, clean headings and answer-first writing serve both.
By sampling. There is no rank tracker for answer engines, so you ask the assistants the questions your customers ask, record whether you appear and in what terms, and repeat on a schedule to see movement.
Quite possibly. Many robots.txt files carry a blanket rule that catches GPTBot and ClaudeBot without anyone intending it. Checking takes a minute and it is the first thing I look at.
How a ai search optimisation (aeo & geo) engagement actually runs, start to finish.
Ask the assistants your customers questions and record where you stand.
Implement the schema graph, llms.txt and crawler permissions.
Restructure key pages so each passage answers cleanly on its own.
Re-run the same prompts and compare against the baseline.
Full Time Contract Available. Tell me about your project, your timeline and your budget, and I’ll reply with an approach and a rate.
Start a project
