For a generation, a client with a legal problem opened a search engine, scanned a page of blue links, and clicked. That habit is quietly being replaced. Increasingly the same client opens an AI assistant, describes their situation in a full sentence, and receives not a list of links but a short, confident answer: a handful of firms, already filtered to their problem. The discovery layer for legal services is changing, and it changes what "being findable" really means.

This is not a distant trend. Search engines now place AI-generated overviews above traditional results for a large share of queries, and a growing number of people treat conversational assistants as their first stop. For law firms the shift raises a practical question and a professional one at the same time, and the two turn out to have the same answer.

From a page of options to a single answer

The old model presented options and let the client choose; ranking and proximity did most of the work. The new model does the choosing and presents a shortlist. The consequence is significant. It is no longer enough to appear somewhere on a results page. A firm must be one of the few a system understands well enough to name. Visibility has narrowed from "listed" to "recommended", and the crowded middle of the results page, where many competent firms used to sit and still pick up work, is exactly the part that an AI answer leaves out.

What this looks like in practice

Consider a client who types into an assistant: "I have a shareholder dispute, the other side has already instructed counsel, and I need a firm that has handled unfair-prejudice petitions in my city." The assistant will try to name two or three firms that visibly match every element of that request. A firm with a clear page on shareholder disputes, named litigators whose experience is stated plainly, and consistent contact and location details is a candidate. A firm whose website says only that it "offers corporate and commercial services", however capable it actually is, is not, because nothing on the page tells the system it fits the specific question. The work existed; the evidence of it did not.

Why this is a conduct issue, not only a marketing one

Whatever an assistant surfaces about a firm is drawn from that firm's own content, profiles and third-party mentions, so the advertising rules apply to it just as they apply to a brochure. The through-line across regimes is the same: communications about a lawyer's services must be truthful and not misleading. The American Bar Association's Model Rule 7.1 states the principle plainly, the Solicitors Regulation Authority's Standards and Regulations require publicity to be accurate, and in India the Bar Council's July 2026 circular on advocates and social media addressed AI directly, prohibiting deepfakes and fabricated citations and requiring disclosure where AI has materially generated content. The lesson for firms is that the accuracy duty which governs a printed advertisement now governs the material an AI reads and repeats, including anything a firm generates with AI itself.

What these systems actually reward

Generative tools do not rank pages in the old sense; they assemble answers from sources they can parse and trust. In practice they favour a few things: clear, consistent information about who a firm is and what it does; corroboration of that information across independent sources; genuine client reviews; and content that answers a specific question directly and accurately. What they discount is ambiguity and unsupported superlatives, which is precisely what the conduct rules discourage. There is a useful alignment here that firms should take some comfort from. The compliant path and the effective path have become the same path, so effort spent satisfying the regulator is no longer effort taken away from being found.

What firms should do

A handful of disciplines cover most of it. Make the firm legible: state the firm's name, practice areas, locations and people consistently everywhere a machine might read them, because ambiguity, not obscurity, is the most common reason an assistant stays silent about a firm. Answer real questions accurately, with clear explanations of the issues clients actually raise, grounded in law rather than adjectives. Put real people forward, because named lawyers with genuine credentials are easier for readers and systems alike to trust than a faceless brand. Treat reviews as evidence, ethically, encouraging honest client feedback and never buying or fabricating it, which both the rules and the platforms penalise. And keep a human accountable, because responsibility for the accuracy of anything published in the firm's name does not pass to software.

The through-line is trust

Strip away the technology and the task is old-fashioned: be genuinely useful, be accurate, be verifiable. Artificial intelligence has simply raised the cost of getting that wrong and the reward for getting it right. Firms that treat their online presence as an extension of their professional reputation, rather than as a billboard, will be the ones the next generation of tools is comfortable recommending. The rest risk a quiet disappearance from a shortlist they never knew they were being considered for.