An AI visibility report.

AI is becoming the front door to search.

For twenty years, getting found meant ranking in Google's list of blue links. For a law firm, plenty of leads still come the familiar way: the Google map pack, the legal directories, reviews, and referrals. What's changing is that a growing share of people now ask a question instead, to ChatGPT, to Google's AI answer at the top of the page, or to a phone assistant, and act on the one answer it gives. For a firm, that adds a new way to get found, or missed: whether an AI names you when someone asks who to call.

Sources: Pew Research (2025); BrightEdge via SQ Magazine (early 2026). Trackers vary, so the 48% is one firm's estimate. The direction is the point.

None of this needs an opinion about whether the shift is good, and it doesn't replace the directories or the map pack yet. It's a fast-growing layer on top, measurable and worth getting ahead of. The practical question for any firm: when someone asks an AI who's a good estate planning attorney around here, does its name come up?

It repeats back whatever it can read.

An AI isn't magic, and it doesn't form opinions about anyone's lawyering. It answers a question one of two ways: from what it absorbed during training, or by reading websites and directory profiles on the spot and summarizing them. Either way, it can only repeat what's written somewhere it can read.

Two things make a firm easy to name: clear text a person reads, and machine-readable details underneath, called structured data, that state the same facts in a form a machine trusts. Birchwood's site has the first and none of the second.

Structured data is a small, hidden block of code in a page. A person never sees it. Google and AI tools read it first, because it states facts as labeled data instead of a sentence they have to interpret. The same information, two formats:

The left is a sentence a machine has to interpret. The right is a fact sheet in the machine's own language, sitting in the page's code, with the attorneys, the bar admissions, and the rating stated as data. Among the firms in this report, the one with the most content carries far more of this than Birchwood, whose site carries only the prose. The next sections show what that produces.

Asked for two names, Birchwood wasn't one.

We asked the two leading AI tools, ChatGPT and Google's AI, the question a growing number of people now ask out loud: name the top two estate planning attorneys around here. Neither one named Birchwood. More people are taking the single answer an AI gives instead of scrolling a list, and when that answer is only two names, not being one of them is the difference between getting the call and never hearing the phone ring.

Why those names and not Birchwood? It came down to things these tools can read, none of them the quality of the lawyering, which AI doesn't measure:

• Published articles. This is the big one for law. Hammond & Pratt has years of plain-English articles and a real FAQ answering the exact questions people search. AI quotes what's written down, and Birchwood hasn't written it down.
• Review and directory signals. Birchwood's 4.9 and its Avvo "Superb" rating are strong, but they live on Google and Avvo, not on its own site as data, so the tools reading websites can't see them.
• Rating. Birchwood's 4.9 is the highest among the established firms in the set, but neither tool weighed it.
• Structured credentials. Bar admissions, law schools, years practicing, the Avvo and Super Lawyers recognition: a machine can only use these if they're marked up as data. None of Birchwood's are, so a machine can't read them.

AI answers shift by date, location, account, and exact wording, so a re-run won't match these word for word; the pattern is what holds.

What the site hands an AI, line by line.

This looks at one thing: how readily an AI tool can pull Birchwood's firm details from its site and trust them. Each line below is the kind we check against the live site, so each is verifiable.

The one item marked present, readable text with named attorneys, is the foundation most firm sites already have. The four marked absent are the signals the AI tools above read, or fail to read, when they assemble an answer. The grade reflects that gap, not the quality of the lawyering.

The five firms, side by side.

We read the live code of the most visible estate and family law firms in the market and paired it with each one's Google rating and review count. The table records what we found.

Structured data and content are read from each site's live code; review counts are each firm's Google total, checked the same week; ratings are approximate. Review volume is low across the board, which is normal for law: people don't post about their estate plan the way they review a home-services shop, so ~140 here carries more weight than the raw number suggests.

Birchwood's 4.9 is the highest among the established firms, and its ~140 reviews are the second-highest volume in the set. The gap is content. Hammond & Pratt, the bigger firm, has spent years building a deep blog and a real FAQ, and it shows: that's a fair loss to concede on content. But even Hammond & Pratt is only running basic business schema, not the Attorney, Person, or FAQPage markup that locks in citations. Everyone else, Birchwood included, has neither. So the schema lane is wide open for whoever builds it first.

FAQs and articles, and who has them.

When an AI answers a specific question, it pulls from pages that answer that question in plain text. Two kinds of page do this work, and for law they're the whole ballgame: they feed Google and the AI tools alike.

Across the same five firms, the answer-content varies:

• NeitherBirchwood: no FAQ and no articles. Nothing published for a tool to summarize, despite the best reputation in the set.
• BothHammond & Pratt: years of plain-English articles and a genuine FAQ. The most answer-content in the market, and the firm AI tends to name.
• FAQStonebridge Family Law: a short FAQ covering common family-law questions. No real article library.
• ThinCedarwood Law: a handful of posts, not enough depth to get cited for the informational searches.
• NeitherOakhaven Estate Law: no FAQ and no articles, with a perfect 5.0 resting on a small review base.

When an AI looks for who to recommend on a specific question, it draws from the firms that have written the answer down. Right now that's mostly Hammond & Pratt, and Birchwood, with the strongest reputation here, has published none of it. It's a difference in what's on the page, not in the lawyering behind it.

The factors behind the differences above.

A few factors account for most of the gap between the firms AI surfaces and the ones it skips. Some live on the site, some on Google and the directories. None requires rebuilding what's there.

Tools in use across small firms, and what each does.

Plain descriptions, no brand names. Each one addresses a specific operational problem. One rule runs through all of it: anything that touches client information has to respect confidentiality and your bar's ethics rules. Don't feed privileged or identifying client facts into a public AI tool, and an AI legal draft always needs an attorney's review before it goes out. The notes mark where a firm of Birchwood's size most often sees fast payback.

No firm uses all of these. For one of Birchwood's size, the typical fit is two or three, not the whole list, and the intake and content pieces are where it usually starts. The foundation, though, is the same for everyone: make the reputation legible to the machines that now do the recommending, without ever handing them a client's confidential facts.