GEO for Professional Services

A growing share of the research stage happens there, though nobody should claim to know the exact proportion for your practice area. What is worth being clear about is the limit of the mechanism: no one controls what ChatGPT, Gemini, Perplexity or Google's AI Overviews say about anything. There is no submission form, no ranking to buy, and any agency promising you a citation is selling something it cannot deliver. What you can influence is whether the material a model would need to name you exists, is unambiguous about who you are and what you do, and appears in the sources those systems draw on. That is the whole of the work. Treat it as additive rather than as a replacement for referrals and search, because the ground is moving and a plan built entirely on it is fragile. The same restructuring makes your pages clearer for human readers.

They share a foundation, and the honest version is that most of the work overlaps. Both need a crawlable site, clear structure and content that answers something. Where they diverge is the target: search optimisation aims at a ranking position, while this aims at being the passage a system extracts and attributes. That changes the writing more than the technical work, meaning answers stated first rather than built up to, passages that make sense lifted out of the page on their own, consistent naming of your firm and its people across every place they appear, and structured data that makes those relationships explicit. A page sitting at the bottom of page one can still be the source that gets quoted. The practical order matters, though. If your site has indexation problems or thin pages, fix those first, because buying this separately while the basics are broken pays twice.

By testing it directly, with the caveats stated up front. We fix a set of the questions your prospects actually ask, run them monthly across the major assistants, and record whether you are named, what was said, and who was named instead. The caveats matter: answers vary between sessions, between users, and between model versions, so a single result proves nothing and only the trend across months means anything. Alongside that, your analytics will show referral visits from the assistants that link out, which is a smaller but harder number. You can run this yourself for free, and we would rather you did than take our word for it, so write down twenty questions, run them once a month, and keep the results in a spreadsheet. What no one can offer is a guarantee that you will be cited, or a lever that makes it happen on demand.