First Mind AI Study: Major Injury Firms Missing From AI Answers in Key U.S. Markets

Lakewood Ranch, United States – July 14, 2026 / First Mind AI /

Many of the most heavily advertised personal injury law firms in the United States are nearly invisible inside AI-generated answers, according to a new study of 280 responses collected from ChatGPT, Gemini, Perplexity, and Grok in the nation’s seven largest legal advertising markets. One in three heavy-advertising firms tracked never appeared in a single answer.

The study, called the PI AI Visibility Index, was conducted by First Mind AI, an answer engine optimization consultancy founded by practicing personal injury attorney Adam Brillhart that helps businesses get recommended by AI assistants such as ChatGPT, Gemini, and Perplexity. The full rankings are published at firstmindai.com/pi-ai-visibility-index. Researchers submitted the same 10 client-discovery questions, such as “best car accident lawyer” and “who should I call after a car accident,” to four AI platforms in Orlando, Atlanta, Tampa, Dallas, Phoenix, Las Vegas, and San Antonio. Each firm’s score is the percentage of its market’s 40 answers that mention the firm.

Key findings

Ad budgets showed almost no relationship to AI visibility. In Dallas, The Law Offices of Frank L. Branson, a trial boutique, appeared in 47.5 percent of AI answers, while Jim Adler, known statewide as The Texas Hammer and the most recognizable injury advertising brand in Texas, appeared in 5 percent. Five more of the Dallas market’s most familiar TV and radio law brands never appeared at all.

In San Antonio, Thomas J. Henry, the second largest legal spot-TV spender in America at roughly 21.5 million dollars per year according to the American Tort Reform Association, trailed the Law Offices of Pat Maloney in AI visibility, 35 percent to 47.5 percent.

Florida belongs to Morgan & Morgan, which appeared in 57.5 percent of Orlando answers and 62.5 percent of Tampa answers, the two most dominant showings in the study. Dan Newlin Injury Attorneys, one of the country’s largest injury firms, appeared in 17.5 percent of Orlando answers. In Atlanta, John Foy & Associates led at 35 percent, and in Phoenix, Lerner & Rowe led at 35 percent.

Las Vegas, the city with more lawyer billboards than any other American market, produced the most fragmented results: no firm exceeded 25 percent, meaning the AI recommendation there is still effectively unclaimed.

The study also documented AI assistants recommending out-of-state firms over local mainstays in three markets, and, in Atlanta, one platform answering an insurance-claim question by listing the insurers Allstate, State Farm, GEICO, and Progressive alongside law firms.

Why it matters

“For decades, injury firms bought discovery: TV, billboards, radio, then Google,” said Brillhart, who also hosts The PI Playbook podcast. “That era is ending in front of us. When an injured person asks an AI assistant who to call, the assistant answers with two or three names, and our data shows those names are usually not the biggest advertisers. Ad budgets do not transfer to the AI layer. Verifiable results, consistent identity, and cited content do.”

Legal queries trigger AI-generated answers at among the highest rates of any industry, and the pattern of consumers discovering firms through AI assistants before searching them on Google is now widely documented in industry research.

Methodology

All data was collected on July 12, 2026 using United States geo-targeting. Ten standardized discovery-stage questions per market were submitted to ChatGPT, Gemini, Perplexity, and Grok, producing 40 answers per market and 280 in total. Brand variants and misspelled domains were merged to the firm entity, and directories, bar associations, insurers, and government resources were excluded from rankings. Scores measure frequency of mention in AI-generated answers and imply nothing about the quality of any firm’s legal services. The complete market tables, question list, and methodology are published at firstmindai.com/pi-ai-visibility-index.