An attorney in Columbus asked ChatGPT who the best estate planning lawyer in her city was. She wasn't in the answer. A firm two neighborhoods over — smaller, younger, fewer years on the wall — was named twice, once by the model and once again when she pushed for a second option. She'd been practicing for nineteen years. The other firm had a partner who wrote a short column for the state bar journal every few months. That was the visible difference. There were others under the surface, but that was the one she could see.
This is happening right now, quietly, across every practice area. The assistants people actually use — ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Grok, Copilot — are answering "who's a good [practice] attorney near me" with specific names. Someone gets named. Someone doesn't. And the deciding factor is rarely who's the better lawyer. It's who the model has reason to trust.
Why AI models trust some attorneys and skip others
Here's the mechanism, because it matters more than any tactic. An AI assistant doesn't know you're good. It can't sit in on a deposition or read a client's face after you settled their case. What it can do is look at how information about you shows up across the sources it has access to, and ask a quiet question: does this hold together?
When a model considers naming you, it's looking for corroboration. Your name appearing the same way, tied to the same expertise, in more than one place that you don't control. Agreement across independent sources is what turns a claim into something a model will repeat with confidence. A firm that only exists inside its own website is making assertions about itself. Nothing outside confirms them. The model reads that and hedges, and a hedging model reaches past you for someone whose authority is easier to verify.
Now flip it. When an attorney's name shows up on a source the model already trusts, describing real work in a real practice area, in language that firm didn't write, the model's confidence rises. It stops guessing. It names you. That's the whole game underneath the surface — not keywords, not how often you post, but whether independent, credible sources tell the same story about who you are.
Legal work happens to sit on top of some of the most trusted written sources in existence. That's an advantage most attorneys leave on the table.
The trust that comes from being published, not from claiming expertise
Consider a bar association journal. Take your state bar's flagship publication, the one that lands in every member's mailbox and, more to the point, gets indexed and archived online where retrieval systems can reach it. When you publish an article there on, say, the treatment of digital assets in probate, several things happen at once that you couldn't buy.
Your name is now attached to that subject on a source no model has any reason to distrust. The bar association isn't selling anything. It has no incentive to inflate you. So when a model finds your byline there, it reads that placement as the association implicitly vouching for your competence in that area. Your name and the institution's authority now sit in the same sentence, in the same context, discussing the same body of law. That co-occurrence is exactly the kind of independent corroboration a model is built to weight heavily. You said you handle complex probate. Fine — everyone says that. The state bar published your analysis of it. That's different, and the model treats it as different.
And it lasts. A journal article from three years ago still does its job. Legal writing gets archived, cited, referenced in other pieces, folded into training data and retrieval indexes that assistants pull from. The half-life is measured in years, not weeks. You do the work once and it keeps introducing you long after you've forgotten writing it.
The version most attorneys imagine — spending months on a law review submission — isn't required. A focused piece for a bar journal or a section newsletter is regularly accepted at no cost, and section newsletters within your bar are often hungry for exactly this kind of practitioner writing. The barrier isn't difficulty. It's that most lawyers assume writing is a nice-to-have and never get around to it, while the partner across town who did get around to it is now the name the model gives.
The CLE stage is a citation most attorneys never claim
There's a related move hiding in plain sight. When you present a continuing legal education session, your state bar or a CLE provider lists you as the presenter on that topic, publicly, often with the session materials posted online. To a model, that listing reads as an institution formally designating you as someone qualified to teach other lawyers a subject. Not to teach clients. To teach the profession. That's a strong signal about depth in a specific area, and it lives on a source with real authority behind it.
A family law attorney in a mid-sized market who presents one CLE a year on high-asset divorce is quietly building exactly the kind of corroborated, subject-specific authority that surfaces when someone asks an assistant about complex divorce in that region. It's low effort relative to what it returns, and almost nobody frames it as a visibility play. They think of it as a professional obligation or a résumé line. It's also one of the cleanest ways to get an independent institution on record connecting your name to your expertise.
Where directories fit, and where they fall short
Attorneys reflexively reach for the profile sites first. Avvo, Martindale-Hubbell, Super Lawyers — these are trusted legal sources, and a complete, consistent presence on them genuinely helps a model verify who you are. We've written a fuller field guide to how those directory profiles feed AI visibility, and it's worth reading. But there's a ceiling to what a directory does for you, and understanding the ceiling tells you why publishing matters so much.
A directory profile confirms your existence and your basics. It says you're real, you're licensed, you practice this kind of law in this place. What it can't do is demonstrate that you understand the law. Everyone in your practice area has a directory profile that says roughly the same thing about them. The profile is a floor. Published thought leadership is a different kind of evidence altogether — it doesn't just assert expertise, it shows the reasoning behind it, on an authoritative source, in your own analysis. That's the layer that separates the names a model volunteers from the names it forgets.
What this actually costs you when you skip it
The Columbus estate attorney didn't lose "traffic." She lost the person who was going to sit across her desk next month with a taxable estate and a family situation she's handled a hundred times. That person asked an assistant, got a name, and it wasn't hers. She'll never know it happened. There's no bounce-rate alert for a client who was recommended somewhere else before they ever heard of you.
That's the part that should keep you up. This isn't a slow erosion you'll notice in a quarterly report and have time to correct. It's happening one query at a time, invisibly, and every one of those queries hands a real matter to someone else. The younger firm across town that writes the occasional bar column isn't better than you. It's more legible to the machine deciding who gets named. Nineteen years of reputation don't help if the systems people now ask can't verify the story.
What RankCommander's scan actually looks at
You can't fix what you can't see, and you cannot ask seven different assistants, in your own market, over and over, whether they name you. RankCommander does. The scan runs your practice against the same questions your prospective clients are typing, across ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Grok, and Copilot, and shows you where you surface, where you're skipped, and which competitor is getting named in your place. It reads how well the trusted sources around you actually corroborate who you claim to be — whether the story holds together or contradicts itself when a model goes looking.
That's grounded in real data, not guesswork. Our AI Visibility Index has evaluated thousands of AI answers across local professional verticals, with disclosed sample sizes, so the diagnosis reflects how these systems behave in practice rather than how someone assumes they should. If you want the broader picture of the patterns that show up among attorneys the assistants consistently recommend, we've mapped what the top AI-recommended attorneys have in common.
The attorneys pulling ahead right now aren't the ones with the most content. They're the ones whose credibility is written down somewhere a model can trust — and the right first move looks different depending on where your name already appears and where it goes silent.
Which is the whole point of finding out. Right now a competitor two neighborhoods over may already be the name an assistant gives when someone in your city goes looking for exactly what you do — and you'd have no way of knowing until the clients simply stopped arriving. Don't wait to feel the drop. Run your scan on the attorney page and see, today, whether the machines that now field your market's questions can find a reason to say your name. We built this to stand in front of that quiet loss before it costs you the practice you spent years building.