AI does not rank your pages; it recommends sources it trusts. For twenty years, visibility meant ranking a page in Google. Now your firm competes to be treated as a credible source worth citing — and credibility, to a machine, is made of five signals most good firms have never deliberately built.
The shift that changes the work
When someone asks ChatGPT, Gemini or Perplexity for a recommendation, they are not after ten blue links. They want a shortlist, a steer, an answer. So your firm is no longer competing for a position on a page; it is competing to be the source the model reaches for. That matters most in professional services, where you sell judgement, trust and risk reduction — precisely the things AI tries to infer from the signals around you. Many credible firms have very weak signals. Here is what authority now looks like.
Signal one — earned mentions beyond your own site
The first question a model effectively asks is: who else says you know your subject? Your website is your base, but AI looks for corroboration from sources that already carry weight — trade media, industry reports, reputable directories, podcasts, association pages, awards, partner sites. You do not need national press. A procurement consultancy wants to appear in procurement contexts; a boutique employment firm in HR and legal ones. The point is not publicity. It is corroboration.
Signal two — content structured to be extracted
AI does not read your site the way a person does. It needs pages that answer questions directly, with clear headings that map to real client prompts, named services, plain definitions and specifics. Many professional-services sites are beautifully written for human impression and useless for machine extraction — all “we help leaders unlock growth,” no “which firm helps UK advisory firms build a lead-generation system.” Write for people; structure for machines. AI cannot cite what it cannot extract.
Signal three — being the same firm everywhere
AI builds confidence from repeated, matching signals. If your name, description, service categories and leadership read differently across your website, LinkedIn, Companies House and directories, you create friction. A loyal referral copes with the ambiguity; a model may not, and it quietly places you in the wrong category or none at all. Consistency is unglamorous, foundational work.
Signal four — depth on a subject, not a mention of it
One blog post is not authority. Authority is depth, repetition and coverage — a connected body of work around one commercially important issue, where each piece strengthens the next and gives the model more context. A thin page rarely signals expertise; a structured cluster does.
Signal five — being referenced by others
The signal most firms miss. It is not enough to say something useful; other trusted sources need to confirm, repeat or link to it. In AI visibility, citations and corroborating references help the model decide a claim is safe to use. Make one core piece of expertise — a framework, a checklist, a briefing — genuinely easy to reference, then put it in front of the partners, journalists and associations who might.
Why this only happens when it has a rhythm
Most firms grasp all this once it is explained. The failure is never awareness; it is execution. The audit gets run, the first article gets written, client delivery takes over, and it stops. Authority signals are not built in a week, which is exactly why the Expert Circle programme treats AI visibility as one of its monthly themes rather than a project. Two people from your firm work the same material — the leader who sees the commercial risk and the marketer who does the work — with an accountability meeting that keeps it moving. When a client asks AI who to call, these are the signals that put your name in the answer. You can see how it works at Expert Circle.
Three things to do this month
- Run a baseline audit, then pick the single topic you most want to be known for.
- Decide your canonical firm details and fix the places they currently disagree.
- Build one citable asset — a framework or checklist — and place it with one external source who might reference it.
The firms that win in AI search will not be the ones with the prettiest websites. They will be the ones with the clearest authority — visible in the right places, structured to be read by machines, consistent, deep and corroborated. That is what it takes to be recommended.
