For law firms of 50 to 150 lawyers

Directories write most of the AI shortlist. Your own pages can earn the rest.

We put 360 questions a client might ask to ChatGPT, Perplexity and Google, five times each, and tracked how often 200 US firms of 29 to 170 lawyers were named. When a question matched a firm’s state and practice area, the firm was named in about 1 answer in 15. When the client described the problem instead of asking for “the best firms,” ChatGPT’s answers cited mostly law firm websites. That is the opening, and we have built for it on a real firm.

  • No charge
  • Three firms a month
  • Yours to keep, whether or not we work together

Inside kraus.law: a narrated walkthrough

Inside kraus.law: the problem, what we built, how it works, and what moved. Narrated by Mark Nelson, with captions. Disclosure below.
Read the transcript

A business owner needs a lawyer. She does what more of your clients now do. She asks an AI. Something like this: “I own a family business and need to plan how it passes to my children. We’re in California. Which law firms should I talk to?” What comes back isn’t ten links. It’s a short list, with sources. If your firm isn’t on that list, that’s the first impression that client gets of your field.

We measured this across 200 firms your size. The lists came back mostly from legal directories, and a firm was named in about one answer in fifteen that fit it. But when a client describes the actual problem, the answer cites law firm websites. That’s the opening.

This is Kraus Law, a two-lawyer business practice in Texas. In September, we rebuilt it as a working legal reference, and it went live on the 29th. The reference isn’t buried in a blog. It’s in the main menu, one click from every page. The core is a glossary: 318 Texas business-law terms, each in plain English, each tied to the statute or the case behind it. A client, a lawyer, or an AI engine can go straight to the term they’re asking about.

Every entry opens with a short answer, two or three sentences, written to be quoted. That’s the passage an AI answer lifts and cites. Then the authority: the rule, the statute, the leading cases. Anyone can check it. And every page carries the lawyer’s name. Nothing went live without his review.

Then original research: every published opinion of the Texas Business Court, coded. Numbers no one else publishes. A case tracker kept current, every row can be cited on its own. And each judge’s own procedures rearranged around what a litigant is about to do.

Here’s what moved. In the 29 days after its release on September 29th, kraus.law drew 612 clicks from Google. In the 29 days before, 38. And when we asked ChatGPT how often the Texas Business Court grants motions to dismiss, the first source it cited was the firm’s own report in two of three runs. It’s early, and it’s one small firm, but it’s the method doing what it should.

Kraus Law is one partner’s practice. A firm of 100 lawyers has a dozen practice groups, each with its own clients’ questions. The firm owns the reference. Each lawyer owns the byline. If you’d like to see where your firm stands, we run the same measurement for three firms a month at no charge. Request it at clarion.studio.

The problem

The first impression now arrives as a short list, with sources.

“I own a family business and need to plan how it passes to my children. We are in California. Which law firms should we talk to?”

One of the 360 questions in the study, word for word.

Picture the person typing that: a real problem, and money to spend on it. What comes back isn’t ten links. It’s a short list of firms, and the sources the answer relied on.

If your firm isn’t on it, or a competitor’s article is the source the answer leans on, that is the first impression that client has of your field.

In-house lawyers, who hire outside firms, are adopting these tools quickly. Among in-house lawyers surveyed in 30 countries, the share actively using generative AI went from 23% in 2024 to 52% in 2025, according to the Association of Corporate Counsel and Everlaw.

Firms your size still win much of their work by referral, and nothing in our study measures how many of your clients ask an AI first. What it measures is what they get when they do.

What 5,400 answers showed

Three findings, and one that cuts against us.

We wrote down every test and registered it publicly before any data came in. Clarion sells website and search work to law firms, so we have an interest in the answer. The results against that interest are reported in full.

Finding one
54%

Ask an AI tool to name firms, and it mostly copies the directories.

Legal directories were 54% of ChatGPT’s cited sources and 60% of Perplexity’s, against 13% of Google’s ordinary results. Three times out of four, when one of our firms was named, the same answer was citing a directory page that lists it. Firms with more Best Law Firms rankings were named more often, even after allowing for size, links and traffic: one standard deviation more rankings went with roughly 1.7 to 3.4 times the odds of being named.

Finding two
1 in 15

Most firms your size rarely appear. Ten firms took 43% of mentions.

When a question fit a firm’s state and practice area, the AI named that firm in about one answer in fifteen. Of the 162 firms with at least one question matching their state and practice, 62 were never named at all. The ten most-named firms took 43% of every mention.

We could not count every firm practicing each area in each state, so this shows how rarely firms appear, not how that compares with chance.

The opening
78%

When the client describes the problem, the answer cites law firm websites.

On those questions, 78% of what ChatGPT cited were law firm websites, against 28% on “best firms in [state]” questions, and directories fell to 11% of its sources. Most of the firms cited were outside our 200.

Found in the data, not registered as a test in advance, and measured on ChatGPT only. Treat it as a lead, not a proven rule.

Technical markup showed no detectable link to being named. We sell that work.

Structured data is widely sold as the fix for AI visibility. We scored every firm’s site before collecting a single answer, and the score showed no detectable link to being named, by AI or by Google. Most sites scored low, so this says little about what a very high score would do.

One firm of about 70 lawyers, assessed from public sources, scored 27 of 100 for markup, in the 87th percentile of the set. It was named in none of 420 AI answers about its own states and practices, and it had no Best Law Firms profile we could find. If someone sells you markup as the answer, ask for their evidence.

The study: 200 US firms of 29 to 170 lawyers, each headcount checked on the firm’s own website. 360 questions a client might ask, across 120 pairs of state and practice area, each put to ChatGPT, Perplexity and Google five times between 24 September and 7 October 2026: 5,400 answers. Registered at osf.io/3rjx2 before collection. The full report, with method, limits and tables, is at clarion.studio/research. A third wave runs through 19 November 2026; these figures will be updated, with dates, once it is analyzed.

The solution

Two kinds of question. Two different fixes.

The study points to two places a firm can act. One you can check this afternoon without us. The other is where a firm can be the source the answer cites, instead of a line in someone else’s list.

“Who are the best firms for…”

Your directory profiles decide this one.

These answers draw most heavily on Chambers, Best Law Firms, Best Lawyers and Super Lawyers. So the first check costs an afternoon: are your profiles there, current, and listing the practices clients actually ask about?

You don’t need us for that. It is in every assessment because it is the least expensive thing to fix.

“I have this problem. Who can help?”

Your own pages can decide this one.

Here the answers cite firm websites that explain the situation in plain terms. A firm with no page on the situation gives the answer nothing of its own to cite.

That is the work: in each practice, the pages a client’s own problem leads to, each with a short answer an AI can quote, the authority behind it, and your lawyer’s name.

How it works

What the work includes.

Directory profiles

Every profile present, current, and matched to the practices clients ask about.

Research by practice

The questions your clients actually ask in each practice group, and who answers them today.

Problem pages and references

Plain answers, glossaries and trackers, each opening with a short answer and the authority behind it, each with a lawyer’s byline.

Your lawyers review, not draft

Nothing goes live without the lawyer’s approval and the firm’s conduct review under its own state’s advertising rules.

Measured monthly

The study’s questions, rerun for your firm against the same 200 firms, so you see what moved and what didn’t.

The firm owns the source

In the firm’s own repository from day one. No license and no platform fee.

“Haven’t we just made each partner easier to take somewhere else?”

It’s the right worry, and it’s why the work is built this way. The firm owns the asset. It lives on the firm’s site, in the firm’s repository, with a named owner, a co-owner, and a written rule for what happens when someone leaves.

The lawyer owns the byline, their name on the work. When a lawyer leaves, the co-owner takes over and the history stays with the firm.

The method, on one small firm

Don’t take our word for it. Open kraus.law.

Kraus Law is a two-attorney Texas business-law practice. In September 2026 we rebuilt its site as a working legal reference. It holds 318 glossary terms in plain English, each with its citation, a coded report on all 131 published opinions of the Texas Business Court, a case tracker kept current, and each judge’s procedures arranged around what a litigant is about to do. It went live at the end of September.

Google clicks per week, kraus.law

Ten weeks, 10 August to 18 October 2026. The new pages went live at the end of September.

Weekly Google clicks to kraus.law Before launch, weekly clicks ranged from 4 to 15. In the launch week, at the end of September, they rose to 156, then 135 and 120 in the two weeks after.
Weekly Google clicks, kraus.law
Week startingClicks
10 Aug15
17 Aug4
24 Aug8
31 Aug7
7 Sep5
14 Sep12
21 Sep12
28 Sep156
5 Oct135
12 Oct120

Source: Google Search Console. Weeks start Monday.

38 to 612Google clicks: the 29 days before the 29 September 2026 release, against the 29 days after
15,201 to 92,727Appearances in Google results over the same two periods
2 of 3ChatGPT runs, tested on 1 and 8 October 2026, where the firm’s own report was the first source cited for “How often does the Texas Business Court grant motions to dismiss?”

Early results for one small firm. Clicks peaked in launch week and were 120 in the most recent week shown. Inquiries are not yet measured, and none of this proves cause. It is consistent with the study’s exploratory finding, and the method is designed to extend practice group by practice group.

What I won’t promise

No rankings, no guarantees. Measured month by month.

AI answers move.

Asked the same question days later, ChatGPT kept about a quarter of its sources. One screenshot, good or bad, proves very little.

Referral still rules.

Firms your size win much of their work by referral. This is about what the referred client finds when they check you out.

Description, not cause.

The study shows what the answers drew on. It doesn’t prove what any one change will do for your firm. That’s why we measure it, firm by firm.

The next step, and it’s a small one

See where your firm stands.

Each month I take three firms of 50 to 150 lawyers and run the study’s questions for their states and practices, at no charge.

  • Your states and practice areas, put through the study’s client questions on ChatGPT, Perplexity and Google, five runs each.
  • Where your firm appears, who appears instead, and which sources the answers lean on.
  • A check of your directory profiles against the practices clients ask about.
  • Your position against the same 200 firms, in a written assessment that is yours whether or not we ever work together.
  1. 1You request it. Two minutes, here.
  2. 2I reply personally to confirm your month and the states and practices to measure.
  3. 3You receive the written assessment, and a 30-minute call to walk through it if you want one.

Built for firms of about 50 to 150 lawyers, because that is who the comparison set represents. If your firm is much smaller or larger, write to me directly and I’ll tell you whether this is the right first step. If what you have in hand is an agency’s website proposal, the proposal review reads it for a firm of any size.

Request an assessment

Three firms a month, in the order requests arrive. If a month is full, you’re offered the next open one.

I read every request myself and reply within three business days. Your details are used to arrange the assessment and nothing else. The privacy page also covers the site’s analytics.

Prefer to write? mark@clarion.studio · (360) 223-6989

Questions

The questions a managing partner asks first.

Is the assessment really free?

Yes. It is how I find the firms the work fits. You owe nothing, and the written assessment is yours to keep. If it shows your firm is well placed, it will say so.

What do you need from us?

Your firm’s name and website, one person to send the assessment to, and, if you like, the states and practice areas that matter most. Everything is measured from public sources. We need no access to your systems, accounts or files.

Who sees the results?

You do. A firm’s assessment is never published or shared by name. The study’s public figures are totals across all 200 firms.

Why only three a month?

Each assessment is run and read by me, not generated and mailed out. Three a month is what that allows.

How sound is the study?

The questions, firms and tests were registered publicly at osf.io/3rjx2 before any data came in. It has real limits. The answers were coded by AI, and no person audited the coding. ChatGPT was reached logged out through a vendor, Perplexity through its API and Google through a results API, not the consumer apps. So far it covers one season. And Clarion, which sells in this market, designed it. It describes what the answers drew on. It does not prove what any change will do.

What does the work cost?

Scope follows the assessment and is sized to the practice groups that need it. Engagements begin at $35,000, and the services page sets out what that covers. The assessment itself is free and carries no obligation.

Do you guarantee results?

No. AI answers change from week to week and nobody controls them. What I can show you is where your firm stands today, what the answers rely on, and, if we work together, what moved, measured the same way every month.

Who is behind this?

Mark Nelson. Clarion Studio is a single-principal studio in Lynden, Washington, founded in 2026. It builds websites for law firms and runs the study described on this page. More on the about page, and the live builds are on the work page.

Directories write most of the shortlist. Your own pages can earn the rest.

Find out which of the two your firm needs first. Three firms a month, no charge, and the assessment is yours to keep.

Request your firm’s assessment