LawFather · Field Study
Denver / Personal Injury / July 2026

When AI recommends a Denver injury lawyer, who does it name?

We asked five AI engines to recommend the best personal injury lawyers in Denver, 1,420 times, across all 30 firms on the first page of Google. Then we measured what actually separates the firms AI names from the firms it ignores. Most of what the marketing industry sells did not move the needle. One thing did.

30firms audited
full first-page census
5AI engines
ChatGPT · Claude · Perplexity
Gemini · Google AI
1,420firm-query-engine
results recorded
1run date
2026‑07‑27
Each mark is one firm. Filled = AI named it for its own core practice areas. Outline = invisible.
Named by AI Not named — even for practice areas the firm handles
Methodology

How we built the dataset

This is a full-population census, not a sample or a poll. We defined the population as the top 30 organic law-firm websites returned by Google for “Denver personal injury attorney,” captured logged out with location set to Denver, with ads and directories excluded. We audited all 30 with the same engine on the same day, so every finding below is a share of the real population, not an estimate.

Unit of analysis
Two units: the firm, and the firm‑query‑engine result. “Named” means the firm’s name appeared in the AI answer.
Query design
Each firm was tested on queries generated from its own detected practice areas, plus generic “best injury lawyer” queries, across a fixed five‑engine panel.
What we measured on-site
Page speed and Core Web Vitals, structured data (schema) by type, Colorado statute citations, content citability, rendering, and crawler access.
Honesty rule
A missing engine result counts as no data, never as “not named.” Every relationship reported is association, not proof. See limits at the end.
Study 01

Research question: when someone asks AI for the best injury lawyer in Denver, how short and how concentrated is the list?

The generic question is owned by a few. Everything past it fragments.

Across the generic “best personal injury lawyer in Denver” answers, AI named 179 different firms. But two firms show up in more than half of those answers, and the rest of the field thins out fast.

Share of generic “best Denver PI lawyer” answers each firm appears in  ·  n = 431 answers
Bachus & Schanker
53%
Zaner Harden Law
52%
Denver Trial Lawyers
46%
Jeremy Rosenthal
40%
Anderson Hemmat
36%
Parker Lipman
24%
…173 more firms
<20%
Read this as: the top five names capture only about 28% of all the times AI named any firm. The other 174 firms split the rest. The generic term is crowded at the very top and enormously long-tailed below it.

What this means if you’re a lawyer reading this

“Best personal injury lawyer in Denver” is the phrase every firm wants to win, and it is the one you are least likely to win. A handful of large, long-established names occupy most of that answer. That is the crowded front door. The rest of this study is about the other doors, which are wide open.

Study 02

Research question: does the same firm win every kind of case, or does the leader change by practice area?

No single firm owns Denver. The leader changes with the question.

The firms that lead the generic term do not automatically lead each practice area. Ask about dog bites, trucking, or brain injury and a different firm is named first, and in a couple of lanes that firm is a focused specialist rather than one of the largest names.

Who AI names first, by practice area  ·  leader and its share of that lane’s answers
Generic “best PI lawyer”
Bachus & Schanker
Dog bites
Ramos Law
Slip & fall
Ramos Law
Bicycle
Ramos Law
Truck / semi
Metier Law Firm
Motorcycle
Anderson Hemmat
Wrongful death
Anderson Hemmat
Brain injury
Schatten Law Firm
Pedestrian
Frickey Law Firm
Product liability
Burg Simpson
Read this as: leadership rotates by lane. Ramos Law leads the premises and cyclist questions, Metier leads trucking, Anderson Hemmat leads motorcycle and wrongful death, and two lanes are led by focused firms — Schatten in brain injury, Frickey in pedestrian. The point is not that small firms beat large ones. It is that the firm leading the generic term is rarely the firm leading a given practice area.

What this means if you’re a lawyer reading this

You do not have to win the whole category to be recommended by AI. You have to be the clearest answer to a specific question. Because the leader changes from one practice area to the next, the firm that dominates “best injury lawyer” does not automatically dominate “best truck accident lawyer” or “best dog bite lawyer.” Those are separate contests, and each one is winnable on its own terms.

Study 03  ·  the finding that matters most

Research question: of all the website fixes the AEO industry recommends, which ones actually track with being named by AI?

We tested the industry’s on-site checklist against reality. Speed tracks visibility. Most of the rest does not.

We measured every firm’s AI visibility against the full checklist of on-site signals the AEO industry sells. Page speed was the strongest, and the only statistically significant, relationship. Generic schema markup and “citability” scoring showed essentially no relationship with who AI actually named.

How each on-site signal relates to AI visibility  ·  rank correlation, n = 30  ·  right = tracks with more visibility
−0.6 (tracks with less)0+0.6 (tracks with more)
Page speed scoresignificant · p = 0.003
 
+.52
Has LegalService schemaweak lean
 
+.31
Cites Colorado statutesweak lean · presence only
 
+.28
Review schema markupno relationship
 
+.09
Attorney schema markupno relationship
 
+.04
“Content citability” scoreno relationship
 
−.02
llms.txt file presentemerging — unsupported here
 
−.11
Slow load time (LCP)significant · slower = less visible
 
−.53
Read this as: the two bars that reach furthest, in opposite directions, are both speed. The most visible ten firms load in about 5 seconds; the least visible ten take about 13. The heavily marketed fixes — schema types, a “citability” score, llms.txt — sit near the center line, meaning they did not separate the visible firms from the invisible ones in this sample.

What this means if you’re a lawyer reading this

Much of the AI-visibility industry sells structured data as the path to getting recommended. On this real Denver market, that claim did not hold: firms with the markup were no more likely to be named than firms without it. Speed, on the other hand, tracked visibility clearly. The most visible ten firms load in about 5 seconds; the least visible ten take about 13.

The practical read is simple. Do not pay for a pile of schema tags and call it an AI strategy. Make the site fast, be the most substantive answer to a specific question, and earn a presence on the sources AI trusts. That is the work that moves visibility.

Study 04  ·  five different games, and they are not worth the same

Research question: is there one AI ranking to win, or five separate ones — and do they matter equally?

The engines barely agree on who to name. And they are not equal in the traffic they send.

There is no single “AI ranking” to conquer. When the same question is asked of all five engines, 82% of the firms named appear on only one of them. Winning is engine by engine. But the engines are worlds apart in how many people they reach, which is what decides where the cases actually come from.

Part 1 — How much the five engines agree  ·  same questions, ChatGPT · Claude · Perplexity · Gemini · Google AI
82%of named firms appear
on only 1 of 5 engines
0.3%of named firms appear
on all 5 engines
~10%average overlap between
any two engines’ lists
Named by 1 engine
82%
by 2 engines
12%
by 3 engines
5%
by 4 engines
2%
by all 5
0.3%
Read this as: the engines are not five windows onto one ranking. They are five different rankings. A firm invisible on ChatGPT may be recommended on Perplexity for the same question. A quarter of all engine-to-engine comparisons shared zero firms in common.
Part 2 — How much website traffic each engine actually sends  ·  share of AI referral traffic, 2026
ChatGPT
74.8%
Gemini
11.6%
Perplexity
7.2%
Copilot
3.5%
Claude
2.6%
Read this as: ChatGPT sends roughly three out of every four visitors that AI refers to websites (SE Ranking, June 2026). Google’s reach is larger still than its referral share suggests, because AI Overviews now appear on an estimated 25–30% of informational Google searches and reach billions of people a month (Similarweb / Presenc, 2026). Perplexity, Copilot, and Claude are winnable, but each sends a small fraction of the traffic.

What this means if you’re a lawyer reading this

You do not have to win “AI.” You have to win the engines that matter, in the right order. Being named is engine-specific, so a single “how am I doing in AI” check is misleading unless it looks at each engine separately. But being named is not equally valuable everywhere. You can rank first on Perplexity and still miss most of the clients, the same way ranking first on Bing means little if you are absent from Google.

The strategy that follows is concrete: win ChatGPT and Google first, because that is where the people and the cases are, then extend to Perplexity, Copilot, and Claude. Knowing how each engine picks its answers, and which one to fight for first, is the whole game.

Study 05

Research question: what share of Denver firms have the basic structured data AI can read, and is it a differentiator?

The basics are table stakes most firms clear. The specifics are where almost everyone is thin.

Nearly every firm already lets AI crawlers in and serves readable pages. That is the floor, and it is not where the gap is. The gap is in the specific, trust-bearing markup: only 5 of 30 firms expose their reviews to AI, and not one exposes an aggregate rating.

Share of the 30 Denver firms with each signal in place
Allows AI crawlers
29/30
Server-rendered
27/30
LegalService schema
16/30
Attorney schema
8/30
Review schema
5/30
LocalBusiness schema
2/30
Aggregate rating
0/30
Read this as: the two grey bars at the top are near-universal, so they cannot explain who wins — everyone has them. Below that, adoption falls off a cliff. And a reminder from Study 03: having this markup did not, by itself, predict visibility here. Treat it as hygiene, not as the growth lever.

What this means if you’re a lawyer reading this

The plumbing on most Denver firm sites is fine. AI can reach and read them. So the honest headline is not “your site is broken.” It is: the readable basics are done, almost nobody has the trust-bearing specifics, and even the firms that do are not winning on markup alone. The next edge is off the website entirely.

LawFather

Your firm was in this study.

Every one of the 30 firms has an individual result: the exact queries where AI named a competitor instead of you, engine by engine. We are happy to walk any Denver firm through its own numbers.

See your firm’s results