Across 21,000 AI answers this week, zero of 22 major markets had all four engines agree on the top firm.
The AI for PI Index™ tracks how personal injury firms appear across AI engines each week - which firms get named, on which engines, in which markets, and where those recommendations come from.
These are the AI for PI Insights™ as of July 20, 2026.
That week’s corpus included:
21,000 AI answers
13,377 distinct firms named
185,780 source citations tracked
259 cities
All 50 states plus D.C.
One note before the findings: this week’s index grew to 13,377 firms - our largest snapshot to date. Because of that expansion, we’re holding the week-over-week firm comparison this issue; comparing the two pools would create a false spike of around 400 new entrants. It returns next week.
Every week, we share one big finding, which firms moved, why it matters to your firm, and a client data point. Let’s dive in.
One Big Finding
The two biggest engines are failing in mirror-image ways.
ChatGPT carries no citation from a firm’s own website in 78.2% of its firm mentions.
Perplexity carries none in just 16.3%.
ChatGPT is recommending firms from what Avvo, Justia, and Martindale say about them - not from what their own websites say. A firm can have the most thorough site in their market and it won’t change what ChatGPT says about them.
Perplexity is doing the opposite. It reads your website and still recommends someone else. About 42% of the time Perplexity cites a firm’s own site, that firm is never named in the answer. It’s pulling the content, not the recommendation.
These are opposite problems. ChatGPT needs your directory profiles to be accurate and current.
Perplexity needs your website to be structured in a way it can use as the basis of a recommendation, not just source material it reads and sets aside.
A firm can address one but not the other and vice versa.
What AI “knows” about you: 71.7% is based on firm marketing. Court records are never cited.
Here is where AI’s recommendations about personal injury firms actually come from this week:
Firm-owned marketing pages: 44.6%
Legal directories: 27.1%
Government and educational sources: about 1% (almost entirely government statutes)
News and press coverage: less than 0.1%
There are no court records in the pool, no bar discipline data, and the entire reference category - everything else - is 28 Wikipedia citations.
What AI “knows” about who is a good personal injury lawyer was written by the lawyers being evaluated and by the directories that sell them placements. AI takes that material and restates it in its own voice, hedging a firm’s own million-dollar recovery claims only about 4-8% of the time.
Who Moved?
The firm tracker: no firm owns a national market - and most local ones are still wide open
The week-over-week movers comparison is on hold this issue as noted above. It returns next week.
What this week’s national picture shows:
Morgan & Morgan is the most-named personal injury firm in the country.
They appear in 7.3% of all 21,000 AI answers this week, are present in every state pool, and show up on all four engines.
They lead Florida at 22.4% and Kentucky at 37.9%. In every other state, they are in the mix but not on top - and they don’t crack the top 25 in 17 of them.
Only 6 of the 13,377 firms named this week cleared even a 1% national appearance rate.
Two-thirds of those 13,377 firms were named exactly once.
The realistic game is local - per-engine presence in your own market. We looked at 22 major markets this week. The four engines didn’t agree on the top firm in a single one of them.
Why It Matters to Your Firm
Your firm shows up on different engines for different reasons - and they require different answers.
Because ChatGPT and Perplexity are reading completely different source libraries, there is no market-wide fix. There are engine-level ones.
Avvo reaches 64% of ChatGPT’s cited answers and 2% of Google AI Overviews’. A firm that owns its Avvo profile has a strong answer for one engine and almost nothing for the other.
One finding this week stands apart from the engine mechanics.
When AI answers a question about a personal injury firm - without the query even mentioning it - 74.7% of those answers volunteer that the consultation is free.
Claude: 96.4%
Perplexity: 85.6%
Google AI Overviews: 80.4%
ChatGPT: 39.6%
It outranks ratings and awards, experience, and contingency fee.
The most-repeated thing AI says about any personal injury firm is an offer every firm in the country makes.
What LawShift Clients Have Seen
SARKISIAN, SARKISIAN & ASSOCIATES, PC ranks #1 in Indiana in the AI for PI Index, appearing in over 30% of AI answers to personal injury queries in their state.
Morris Law went from 0% to 20% AI visibility in 45 days and signed a six-figure slip-and-fall case.
The firms winning AI recommendations right now are the ones each engine finds easiest to read and cite.
Everything in this issue happens before a firm ever hears from the injured person. What happens in that window, from injury to intake, is what we call The Pre-Intake Problem™ - and no one is serving it.
Want to know what AI is telling injured people in your market?
Request a Pre-Intake Revenue Strategy discussion. Thirty minutes on your market and your revenue - who AI recommends, who it leaves out, and what that means for the cases you sign. No pitch. No strings attached.
Schedule your Pre-Intake Revenue Strategy Discussion at LawShift.ai.
Until next week,
Nick Kringas
Lydia Flocchini
Injured-First. Always.
