How law firms get found, and cited, by AI search
By Mohammed Buhariwala · 2026-08-05
The search a firm never sees
A prospective client with a problem no longer starts on a search results page and scrolls. Increasingly they ask an AI assistant a plain question, "who's a good firm for X near me?", and act on the two or three names it returns. That entire conversation happens off your website. If your firm isn't in the answer, you were never in the running, and you'll never see the enquiry you didn't get.
This isn't a fringe behaviour any more. Analysts estimate that a meaningful and fast-growing share of informational queries, into the double digits, now resolve inside an AI answer rather than a classic list of blue links, and Gartner has predicted that traditional search volume will fall sharply as AI assistants absorb it. For a profession where a single new instruction can be worth thousands, being absent from that answer is expensive in a way that never shows up in your analytics.
Why most firms are invisible to AI
AI models don't recommend a firm because its website looks polished. They recommend firms they can resolve as a clear, corroborated entity, and whose expertise is stated in a way a machine can extract cleanly and quote with confidence. Most firm websites fail on both counts. Services are described in brochure language a model can't turn into a specific answer. The firm exists as an entity only on its own domain, with nothing independent to cross-check. So even when the model knows you exist, it has no structured reason to put you on a short list of names it's willing to stake an answer on.
There's a second, less obvious problem: the sources these engines actually trust. Independent studies of AI citations have found remarkably little overlap between what different engines cite, one analysis found only around 11% of domains were cited by both ChatGPT and Perplexity, and that community and reference sources like Reddit and Wikipedia are cited far more often than most brands expect. A firm that has invested only in its own polished site has invested in the one source AI engines weight least.
What the research says actually works
The most rigorous public work on this comes from a Princeton-led study, "Generative Engine Optimization" (published at ACM SIGKDD 2024), which tested specific content changes across thousands of queries. The techniques that moved AI visibility most were not tricks. Adding citations to authoritative sources, adding relevant statistics, and adding direct quotations each lifted a page's visibility in AI answers by roughly 30 to 40%. Vague, unsourced, adjective-heavy copy, the default register of most professional-services websites, did the opposite.
The lesson is uncomfortable for a profession trained to sound authoritative through tone. AI engines don't reward tone. They reward specificity they can lift verbatim: a clear claim, a real number, a named source. The firms that get cited write like they're being quoted, because they are.
Trust is now a machine-readable property
The trust that wins instructions has always been built on reputation, but reputation is now something a machine reads as much as a human does. The same signals that make a firm credible to a client, being known, being referenced by others, being clear about who you help and why you're qualified, are the signals that let an AI model resolve you as a distinct, trustworthy entity worth naming.
This aligns with what buyers themselves say. Edelman and LinkedIn's long-running B2B research has repeatedly found that a majority of decision-makers, around 73% in recent years, judge an organisation's capability more from its thought leadership than from its marketing materials. AI engines are, in effect, automating that same judgment at scale: they sample what you've published and what others say about you, and decide whether you're a name worth surfacing.
Getting found is only half of it
Visibility creates enquiries. Manual intake loses them. It's common to see a firm win the hard part, showing up where buyers now look, and then quietly leak those enquiries through inboxes and callback promises that don't happen fast enough. The research on response speed is brutal: the well-known Harvard Business Review study "The Short Life of Online Sales Leads" found that firms which followed up within an hour were far more likely to have a meaningful conversation with a decision-maker than those who waited even sixty minutes longer.
The firms that win treat visibility and capture as one system. Being the name the AI gives is the top of it; catching and answering every enquiry that name generates, instantly, is the bottom. We keep the exact build proprietary, but the principle is simple, and it's where we start: give the machines a clear, corroborated reason to name you, then make sure not a single enquiry that follows is lost.
Get found by AI, then capture what it sends you. Do one without the other and you're either invisible or leaking. See the legal systems we build.
Sources
mohammed@scaleitupmedia.co.uk · LinkedIn · YouTube