Law firms, accountants, financial advisers and consultancies get chosen on trust, and the first place that trust gets tested is now an AI answer. Peak Answer measures whether you are named in it, on the questions that bring in work.
This page is for you if
Professional services sit at an unusual intersection: the purchase is high-trust and considered, the buyer is often researching something they do not understand, and the answer they get from an AI is frequently the first short list they ever see.
That first list matters more here than in most categories, because the buyer has no independent way to judge the options. In a product category they can compare specifications. Choosing a solicitor after a car accident, or an accountant for a first VAT return, they are choosing on the strength of what they were told, and increasingly what they were told came from a model.
We ran the measurement across fourteen markets, three of which are professional services, and the results are more encouraging than the category usually expects.
In personal injury law — a market everyone assumes is owned by the largest advertising budgets — the entry bar came out at 39 and the firm we tested cleared it on three of five questions. On four of the five, a site scoring below the bar was cited anyway. In paid search that market is brutal. In AI answers it has a long tail.
Fractional CFO services produced a bar of 26 in the thinnest citation set we measured: fewer sources cited per question, which means each one carries more weight. Meta ads agencies in Australia produced the opposite result, a bar of 49, and taught us something important: adding a country to a question does not make the answer local. The engines treated the geography as a topic, not a filter.
Across every professional services market we measured, content share ran at 100%. Every single source the engines cited was an article about the subject rather than a firm's own website.
That is the most actionable finding we have for this category, and it points the work somewhere specific. Rewriting your practice-area pages will not move this. Being the firm quoted in the guide, the explainer, the state-specific article, the trade publication piece — that is the mechanism, and it is the only one we observed operating.
It also suits the way professional firms already build reputation. Contributing expertise to publications is not a new activity for a partner; it is what speaking at conferences and writing for trade press already was. What changes is that the audience now includes a model.
The product does not invent claims. Generated content is written from measured gaps and the sources the engines actually cited, and every draft is yours to edit or reject before anything is published. Nothing goes live without approval, and nothing is posted to a third-party platform automatically.
Where the product is uncertain it says so. An article's effect is reported as measured or as too thin to call, never assumed, and the measurement itself carries its sample size.
Week one: the questions your prospective clients ask get measured across six engines, and you find out which ones name a competitor firm instead of you.
After that, the work is ranked. Usually it begins with the technical faults stopping engines reading your site at all, then the explainer content aimed at the specific questions you lose, then getting into the publications the engines are already quoting for your area of practice — with the outreach drafted for you.
Nothing is published without your approval, no claims are generated that are not grounded in the measurement, and no third-party posting happens automatically. The compliance question is the same one you apply to any marketing copy: somebody has to read it before it goes out, and the product is built on that assumption.
Referrals still matter. What has changed is that a referred prospect now frequently checks the recommendation with an AI before calling, and what that AI says about you — or does not say — is part of the decision.
Yes, with the caveat our research turned up: the engines do not treat a geographic qualifier as a filter on who can be named. You will be measured against national content as well as local, so plan for that rather than assuming the locality protects you.
In the professional services markets we measured, bars ranged from 26 to 49. That is a meaningful spread, and the measurement tells you which end of it you are at in the first week rather than after six months of work.
Seven days, no card. Peak Answer measures your first questions on the day you sign up.
Start a 7 day trial