The current conversation
Google’s guidance for generative AI search says established SEO practices remain relevant. It also says Google does not need an llms.txt file or special AI markup to include a page in its search experiences. Being technically eligible does not guarantee indexing or visibility. That is a useful starting point when deciding which changes deserve your budget.
Start with the questions behind the enquiry
Our practical recommendation is to review the questions your team answers before a customer buys. A publisher might need to explain what a package includes. A training institute might need to clarify course duration and entry requirements. Put those answers on the relevant service page, with enough context for a visitor to make a decision. Avoid scattering the same generic paragraph across many locations or industry pages.
Make your claims easy to assess
Describe who delivers the work, what is included and where the information comes from. Use approved project examples when available. If a result cannot be verified, explain the scope of the work instead of inventing a number. This creates a more useful page for prospective customers and a clearer brief for the people maintaining the site.
Connect the information to a next step
A well-written answer should sit within a usable journey. Link a service explanation to the relevant portfolio example and enquiry form. Check headings, mobile layouts, descriptive titles and internal links. Keep key information in readable page content rather than placing everything inside an image.
What to review next month
Choose a small set of meaningful pages and note the current questions, impressions and enquiries available in your reporting. Improve content where visitors still lack information, then review changes over time. Treat search and AI visibility as an ongoing discovery process. Do not buy a promise that one file, one keyword list or one score will guarantee mentions.
Turn a generic service page into a useful decision page
Consider a hypothetical agency page that says it provides innovative solutions, strong results and personalised support. Those phrases reveal little about the work. A stronger page could explain which campaigns are managed, what assets the client must supply, who handles lead follow-up and what will be reviewed after launch. It can show a relevant project and explain its scope without exaggerating the result.
Make a list of unanswered questions for each priority page. Group them by decision: suitability, deliverables, process, evidence and next step. Ask the relevant team member to review the answers. This exercise often exposes gaps in the actual offering as well as gaps in the copy, which makes it useful beyond search visibility.
A practical content maintenance process
Assign an owner to important service information. Record when a page was reviewed, which facts changed and which supporting assets are approved for use. When an offer changes, check related articles, forms and internal links so visitors do not encounter conflicting versions.
Choose a small number of substantive improvements for each review cycle. An original explanation from the team doing the work may be more useful than another broad article repeating familiar advice. Keep the information readable for the intended customer and use normal, descriptive language for headings and links.
What not to promise your stakeholders
Do not equate publishing a long article with earning visibility in AI answers. Do not present an appearance in one personalised search session as a stable ranking result. Explain which observations came from a tool, which came from actual enquiries and which remain hypotheses. A credible search programme gives the business a way to learn, rather than a collection of claims it cannot verify.
Sources & further reading
Google Search Central: guidance for generative AI search ↗Practical frameworks are First Thought editorial recommendations. Hypothetical examples are not client results.
Let’s talk growth