The most dangerous situation in the professional service industry was not to be robbed by rivals, but to be removed from the list by AI before the client decided who to look for. When a new finance minister opens ChatGPT and asks "What accounting firms in Taipei are good at cross-border equity structures," AI will just spit out three to five names with reasons for recommendation. Without you on the list, clients won't know you exist, and you'll never see this loss backstage. This is not the future. This is what's happening.
Specialized service acquisition. It was a question-and-answer game.
Lawyers, accountants, management consultants, services such as high-price, high-trust, door-to-door services have never been settled by a key word commercial. The path of the past was to ask friends, to ask their peers, and to ask the community: "Who was your last lawyer dealing with the takeover?" "The recommendation chain determines who is on the list. The search is about automating, scaling and answering this "question" chain more quickly than most people do.
The difference is that the friend who used to endorse you recommended you because he really worked together. AI didn't work with you. It only read public data. Who it recommends, and what it recommends, depends entirely on how clear and easily readable your professional content is online. A top-notch firm that does not have a readable asset does not exist in the eyes of AI.
AI decides who to recommend.
In response to a question like "Who to find," AI's engine is not randomly chosen or much more advertising. What it is doing is closer to a dedicated assistant: finding out from the public data that "there is concrete evidence of professionalism and clarity". We've seen three decisive conditions over and over again as we're tracking professional service clients.
- Whether the professional depth is readable: the type of case you have dealt with, your interpretation of a particular law or code, has become a well-articulated article, not locked in your head or only closed-door briefing.
- Whether your identity and profile are clear: AI needs to know what you are — lawyers who specialize in labor disputes, tax accountants who serve foreign businesses, counsel for organizational transformation. It's very difficult to classify a vaguely located firm, and it's difficult to get out.
- Whether third-party language is the same: media coverage, business records, speeches, client comments, how you are described by sources other than you, will strongly affect AI's confidence in you.
"We're good enough," why did it start to fail?
A lot of our senior partners' instincts are that we depend on relationships and referrals, not content. That was perfectly correct five years ago, and it is only half right now. The referral remains important, but the first step is moving. One of your potential clients, recommended by a friend, will search your name, your office, and cross-check it. If this round is confirmed to fall on the AI search, and AI has little to say about you, the original accent is diluted into a question mark.
More crucial is the source of new clients. In the past, a decade-long network of people was built up, and now a start-up, with no connections in Taipei, is going to outsource the acquisition policy directly to AI. This group of people is the fastest-growing, least-funded and least-established customer. They don't ask your old clients, they ask for models. You're absent from this conversation, which is tantamount to actively handing over a new generation of clients to their peers who want to run content.
The firm is supposed to be established as a "extractable professional asset."
The solution is not to write a bunch of SEOs into hydrology, but to lower the profile. The AI engine's preference is the content that can be extracted cleanly and answered directly. For a professional service, this means writing out the real problems you're dealing with on a daily basis on behalf of your clients, in a structured article: a context, a rule of law or a code, a judgment and a recommendation. The content, AI, can read, quote and be trusted by human clients.
Let's get this straight.
There's one thing worth doing before you write: actually ask. Open a few mainstream AI engines and use your ideal client's voice to ask, "Who should deal with a certain problem?" to see if you're on the list, how AI describes you, whether you're tied to the wrong professional domain. This action can be done in 10 minutes, but it is often the first time that partners realize that AI describes them as either blank, late or not.
We thought we were famous in the circle, and it wasn't until after a round of interviews with AI that we found out that it recommended all of our competitors, and that the reason for that was to start. That moment was really tense.— A Taipei management consultant partner.
From being displaced to being recommended.
The moat of professional services has always been trust, and trust needs to be seen to be established. AI's search did not change what clients wanted, it just changed the way they looked for and who was entitled to appear in the answer. The good news of this is that most of the industry has not yet started running what can be quoted, and that the business office that is now in operation is just as well ahead of a crowded track. I want to know how AI describes your firm, which one is missing on the list, and which one is scheduled for a 30-minute GEO diagnosis, and we'll actually show you where you are and the gaps in the AI answer.
