By the CW Team
August 2026. 5 minute read.
When someone asks an AI assistant to recommend a lawyer, the assistant does not guess. It reads. In the seconds before it writes its answer, it is drawing on directories, profiles, reviews, articles, and your own website, then deciding which names it can stand behind. Firms that appear in those answers tend to have the same five things in order. Firms that never appear are usually missing at least two of them.
1. Listings that agree with each other
Assistants lean on legal directories: Avvo, Justia, FindLaw, your Google Business Profile, the state bar. These are the sources they cite most when naming firms. Two problems get firms silently dropped. Thin listings, where the profile exists but says almost nothing. And conflicting listings, where the address on Avvo disagrees with Google, or a departed partner still headlines the state bar entry. When sources disagree, the safest move for the machine is to name a firm whose record is clean instead.
The fix is unglamorous and effective: claim every listing, complete every field, and make them match.
2. Attorney bios that read like results
Engines quote specifics. A bio that says "dedicated advocate with a passion for justice" gives the machine nothing to repeat. A bio that says the attorney has tried motor vehicle cases for eighteen years and recovered a specific eight figure verdict gives it exactly the kind of sentence answers are built from.
Read your bios the way a machine would: what facts could be lifted from this page and placed in a recommendation? If the honest answer is none, the bios are decoration.
3. Pages that answer real client questions
People ask assistants process questions before they ask for names. How long does a settlement take. What should I do after a crash. Is this insurance offer fair. When an engine answers those, it cites the pages that answered them best, and the firms behind those pages inherit the credibility.
Most firm websites have practice area pages and no answers. A single thorough guide to the settlement timeline in your state can start appearing as a cited source within weeks, and a firm that gets cited on the process questions gets named on the "who should I hire" questions.
4. Reviews with a pulse
Assistants weigh reputation signals the way a careful referral source would: how many reviews, how recent, what they say, and whether the firm responds. A wall of five year old reviews reads as a firm past its peak. A steady drip of recent, specific reviews reads as a practice in motion. You cannot fake this one, and you should not try. You can build the habit of asking satisfied clients at the moment the matter closes.
5. Being named by sources the machine already trusts
Citations compound. When local news covers your verdict, when a legal publication quotes your partner, when a bar association lists your award, the engines absorb those mentions and treat your name as safer to recommend. Every earned mention makes the next recommendation more likely. Firms with real results usually have more of this material than they have ever published. The gap is not the record. The gap is that nobody wrote it down where machines read.
How to know where you stand
You could check all five by hand for your firm and every competitor, across every engine, every week. Nobody does, which is why nobody knows their own visibility.
Casewell does it as a service. Our scan asks the major AI engines more than 500 real client questions weekly, records which firms get named and which sources get cited, and turns the gaps into a monthly plan. Then we do the fixes: the content drafted, the listings corrected, your job is a yes.
The first scan tells you which of these five is costing you answers right now.
See your visibility.
See what the assistants say about your firm.
Or read what a Casewell engagement includes and what it costs.