
👋 Good morning. Chris Dreyer here. A New York firm paid an agency $300,000 a month for personal injury leads and rejected 77% of them. Now it's suing for fraud, and the complaint is public. Yikes!
Then new data on what AI recommends to a car accident victim looking for a lawyer. Four of the top five names it gives are directories. The study also flags two tactics agencies are aggressively selling right now that do nothing for this audience.
And Chris Earley came back on the show to make a case I've been chewing on since we taped. He thinks most firms chase the wrong kind of growth, and he backs his version with 20 years and 1,007 reviews. Let's get into it.
📆 [SAVE THE DATE: Oct. 4-6]
PIMCON, the Personal Injury Mastermind Conference, is where personal injury attorneys master marketing, intake, and explosive firm growth. It's built for decision-makers who want sharper strategy, better execution, and real-world results long after the sessions end.
October 4–6 • Scottsdale, AZ
You get three days with the operators behind the country's fastest-growing PI firms, and the playbooks that got them there. Lock in your spot now.

💡ONE BIG IDEA
Stop Trying to Win a State. Win Five Square Miles.

Ask PI firm owners what growth looks like and they'll describe a bigger map.
Another county. A second office. A campaign that runs statewide instead of citywide. Every year the territory gets wider, and the same marketing budget stretches thinner.
Listen, you're not making it in this business without healthy ambition. Every firm should want more market.
But Chris Earley, longtime Boston PI attorney, sees it differently. He credits a mentor for the shift in mindset.
"It's one thing to try to dominate a state," Earley says. "It's another thing to try to dominate five square miles."
His logic? The same dollar buys more repetition in a smaller area, and he wants to be everywhere inside it.
A firm competing statewide bids against every firm in the state. Earley bids against the handful who operate in a five-mile grid. Then he kills them with just six words.
"Call Earley before it's too late."
A marketing assistant gifted him that tagline four or five years ago and then left for a better job. He trademarked it, and has followed this simple rule since then: When something lands, double down and go to town on it.
Now it's on the billboards, it's on social media, and soon you'll hear it on the radio. All of it within the same five-mile stretch of Boston.
Not just paid channels, mind you.
The cheapest repetition he buys isn't paid at all. He sponsors local organizations, then shows up and promotes them on social, and earns a ton of mentions in return.
The payoff? Those mentions drive SEO, and he told me he's noticed they feed AI platforms too.
None of that happened fast. He's at 1,007 reviews, which took years of asking every client, and it took him 20 years to reach his first seven-figure month.
Let's map this onto your own market…
Your version won't look like Earley's. In his words: "If I was starting out, I wouldn't get a billboard."
Instead, he'd post free on TikTok, LinkedIn, and YouTube shorts until something landed, then put money behind the winner.
And even then, he thinks PI firms should adopt an approach to marketing spend his coach Mickey Love taught him:
What's the best use of your next dollar? Answer it with two numbers: What it costs you to sign a case, and how many of the leads you actually wanted became clients.
Lock that in. Then pick the ground your firm means to own and buy repetition there, rather than spreading reach across a territory you'll never lead.
Four moves get you there:
Draw the boundary and name the streets. Not the metro, not the county, not the DMA. The actual area, with roads on every side of it. Here's the test: If you can't say out loud which streets bound the ground you're trying to own, you haven't picked it yet.
Put your spend and your posts inside the boundary. Tighten the geo radius first, then push what you save into frequency in the smaller area. If you're not buying media yet, the same rule holds: Post into those few miles rather than the whole metro, and let something land before you pay to amplify it. When you do buy, a small footprint negotiates well. Earley finds that it's easier to negotiate a low price when you bundle advertising, like billboards, in a smaller geographic area than across an entire city, with a Boston highway board closer to $10,000 a month and smaller areas about half that.
Move the sponsorships out of the donation budget. If your firm already writes checks to local teams and charities, the check isn't the work. Show up, photograph it, and post about the organization by name. Then test it: Search the last team you sponsored and see whether your firm turns up anywhere in the results.
Prove you convert inside the boundary before you widen it. Earley holds his intake team to 93.5% conversion on the leads he wants, everybody there knows the number, and he meets that team every single day. If your firm can't convert inside five miles, another county won't fix it. It'll just cost more to find out.
None of this means giving up on a bigger market. Own five square miles and the sixth is easier, then the tenth.

♟️STEAL THIS PLAYBOOK
Grade Every Video on the Job You Gave It

Total views is the wrong metric to gauge the success of your firm's video strategy.
Views are comforting. That number kinda, sorta lets you off the hook, as long as it's higher than the last video you posted.
In Wyzowl's 2026 survey, 67% of video marketers measure the return on video by views, and 63% lean on engagement, meaning likes, shares, and reposts.
Now don't get me wrong…
Those numbers aren't useless. When you build a video for reach, reach is the right measure.
But here's the problem:
Your firm doesn't make one kind of video. You run a CTV brand spot. You post a client testimonial. You publish a "what to do after a truck wreck" explainer on YouTube. You send a short clip to a lead who stopped answering the phone.
Four videos, four different jobs, and the same view count grading all four.
That's like grading your trial team, your intake staff, and your bookkeeper on who got the most calls.
Paige Peterson, who writes the Video Brief newsletter on video strategy, put it this way: "Views can tell you whether a video was seen. They can't tell you whether it worked."
So before you publish the next one, do this:
Write the job down before the video runs. One sentence, and it has to be specific: This spot exists to get noticed in Tampa, this testimonial exists to make a truck-wreck client believe you handle cases like hers, this explainer exists to get found at 2 a.m. If nobody at the firm can say the job out loud, the video isn't ready to publish.
Pick the number that matches the job. A brand spot gets cost per view. A testimonial gets completion rate and saves. A practice-area explainer gets watch time, search traffic, and clicks through to intake. Search traffic just got easier to pull, since Google's platform properties in Search Console now shows which queries send people to your YouTube, Instagram, and TikTok posts. A clip sent to a lead who went quiet gets one number, and it's calls booked. Decide in advance and nobody grades the video on whatever it happened to win.
Stop ranking videos you built for different jobs. Put a CTV spot and a testimonial on the same report and the spot wins nearly every month, because a spot bought for reach pulls more views than a two-minute story about somebody's spine surgery. Rank the spot against your last spot, and the testimonial against your last testimonial.
Ask intake which video the caller mentions. Add one line to the script: Did anything you watched or saw bring you here? The answers come back messy, and they're still the only numbers in the building that tie a video to a signed case.
Do that for a quarter and your video report answers a better question: Did each of these do the job you gave it?
🔗 Video Brief →

📰 TOP OF THE NEWS
Law Firm Sues Marketing Agency Over Bad Personal Injury Leads

A New York personal injury firm is suing its marketing agency for fraud over $1.15 million in fees.
Sanders Aronova Grossman Woycik Viener and Kalant, a Garden City firm, sued Elite Accident Help Corp. on July 21 in the Eastern District of New York. The complaint alleges fraud, fraudulent inducement, and breach of contract, and asks the court to void the agreement's arbitration clause.
The complaint alleges the following:
The deal priced a qualified lead at about $5,000. According to the complaint, Elite represented it could deliver qualified leads at roughly that cost, which put the $300,000 monthly fee at about 60 qualified leads a month. The firm says it has paid no less than $1,150,000 since September 2025.
Rejections ran 77% and climbed in the months around signing. Of 265 leads delivered between October 2025 and April 2026, the firm says it rejected 204. Its month-by-month figures climbed from 62.5% in October to 86.4% in February, dipping to 85.4% in March, with the Independent Contractor Agreement taking effect February 1. Of everything delivered, the complaint says 57 matters remain active and four have settled.
The agreement counted a lead as qualified only if the client stayed to the end. It defined a qualified lead as a matter the firm accepted and the client carried through the conclusion of the case, and it obligated Elite to replace rejected leads within 30 days. The firm says Elite did not supply those replacements. The same agreement required the firm to indemnify Elite over Telephone Consumer Protection Act (TCPA) and other marketing complaints, which the complaint calls a one-sided allocation of risk.
The firm is asking the court to void the arbitration clause. The agreement routes disputes to JAMS mediation and then to a single arbitrator with no appeal, with costs split evenly. The firm argues that clause was procured by the same fraud as the rest of the agreement, so no enforceable agreement to arbitrate ever came into existence, and argues in the alternative that it is unconscionable.
The firm sent a notice of material breach on June 5 demanding 143 additional qualified leads within 15 days. It says nothing was cured.
🔗 CourtListener →
AI Sends Injured People to Directories, Not Law Firms

A new study breaks out what AI recommends to someone hurt in a car accident, and it's mostly not law firms.
Ben Wills, founder of the AI search research site OppAlerts, has now published the consumer legal services slice of his 403,000-prompt analysis, which covers 100 industries. It tracks 1,342 domains across 11 buyer types, and one of them is an injured commuter: a car accident victim juggling medical bills and missed paychecks who needs a fast, local firm.
For that buyer, the five domains AI recommends most are Avvo, Nolo, FindLaw, Justia, and Morgan & Morgan. Avvo scores more than three times Morgan & Morgan, and Morgan & Morgan is the only law firm in the group.
What tracks with getting recommended to that buyer:
Links from the pages that already rank. The dominant factor by a wide margin, and it explains about a quarter of the variation on its own. Current rankings for injury questions matter more in AI citations than anything else Wills measured.
Reddit. Posts and comments both rank high here. Across the whole consumer legal category, Reddit comments are the single strongest predictor, at more than double the all-industry figure.
Backlink authority, still. PageRank, its history, and raw backlink counts all register as strong for this buyer.
Right now, agencies aggressively sell two tactics that do nothing for this audience. Homepage keyword relevance comes in near zero, and a Wikidata entry at zero. Both carry weight in other industries.
Read all of it as correlation, not causation. Wills says so himself, and he warns that any factor with fewer than 20 observations deserves caution, which applies to a few rows in the injured-commuter table.
His broader point holds either way: The evidence AI draws on for injury questions sits mostly on other people's websites.
🔗 OppAlerts →

🚀 QUICK HITS
Morgan & Morgan's Newest Ad Stars a Merman: Morgan & Morgan released "Merman," the latest spot in its long-running "Power Move" campaign, DesignRush reported. A lifeguard watches a public pool with total authority until the camera reveals his tail. The spot closes on the campaign's tagline: A merman lifeguard is a power move, and so is hiring the country's largest personal injury firm. Chief Creative Officer Carlos Wigle said the team "wanted to create something that would make people stop scrolling, lean in, and laugh." It runs nationally through 2026 across broadcast, digital, audio, and out-of-home.
A New Orleans Attorney Buys Ads on His Own Name, Not "Personal Injury Lawyer": Lee Rudin explained on LinkedIn why he runs paid campaigns on his name, his firm's name, and his "Get Rude" tagline rather than bidding on category keywords. SEO agencies keep pitching him page one on Google, he wrote, without pricing in the competition. The firms in his market have been at it 30 years against his two, and he isn't going to run a race he can't win.
A Slack Killer Where Lawyers Work Directly With Their AI Agents? Jack Dorsey launched Buzz on July 21, a workplace group chat platform that puts people and their AI agents in the same conversations, TechCrunch reported. Built by Dorsey's company Block, it looks like Slack, works with any AI model, and is open source, so a team can build and deploy its own features. Buzz describes itself as in early stages, and TechCrunch advised against moving a team onto it yet. The desktop app is free for macOS, Windows, and Linux.
Tennessee Tells a Jury Meta Disregarded Its Own Research on Teen Harm: Opening statements in the state's seven-week trial began Monday in Nashville, the Guardian reported. Tom Cartmell, a lawyer for the state, showed jurors a 2017 internal document in which Meta product managers wrote that notifications and infinite scroll "are inherently at odds with well-being" and said the company needed to warn the public. "This warning never came," Cartmell said. Meta's lawyer said the same documents show a company looking for problems so it can fix them. The case follows a March verdict in New Mexico, where a jury ordered $375 million in civil penalties.
A Senate Bill Would Make Staged Truck Crashes a Federal Crime and Reach the Attorneys Involved: Sen. Ashley Moody, R-Fla., introduced the Staged Accident Fraud Prevention Act on July 22. It would make intentionally staging a collision with a commercial motor vehicle a federal crime, penalize anyone who organizes or takes part in one, and hold attorneys, physicians, and other co-conspirators accountable if they knowingly participate. Participants and organizers would face up to 20 years in federal prison, with a 20-year minimum where a crash causes serious bodily injury or death, FreightWaves reported. Trucking, insurance, and tort reform groups back the bill.

💯 NUMBER TO NOTE

A federal jury in Boston awarded $88 million to an Alabama man and his wife.
Larry Patterson received a Covidien hernia mesh implant in 2017 and says it caused a bowel obstruction that required another operation three years later. After a three-week trial, jurors found that Medtronic's Covidien unit failed to warn doctors about the risks of the product, Reuters reported. They awarded no punitive damages and did not find that Covidien committed fraud.
Patterson's case is the first of thousands against Medtronic to reach a jury. Medtronic says it will challenge the verdict in post-trial motions and on appeal.
The plaintiff's lawyer calls it by far the largest mesh verdict yet. Timothy O'Brien said it exceeds the results in more than a dozen hernia and transvaginal mesh cases already tried. "This is a very substantial verdict, but it reflects very substantial damages," he said.
About 10,350 people are pursuing the same claims. Roughly 7,450 sit in Massachusetts state court, and another 2,400 are consolidated before one federal judge in Boston. The claims run across Covidien's Parietex, ProGrip, and Symbotex lines.
Medtronic is the last major defendant standing. Units of Becton Dickinson and Johnson & Johnson largely settled their hernia mesh cases. Medtronic says its products "have been used safely and effectively in millions of patients."
For a firm carrying mesh cases, this is the first jury number in the litigation, and 10,350 claims sit behind it.
🔗 Reuters →

🎙️ FROM THE POD
Jim Andresen on Who Owns Your Firm's Data

Where your systems don't talk, someone at your firm copies and pastes.
Jim Andresen has run personal injury operations for years. He's the COO of TORKLAW, which practices across multiple states, and he built LawWorks, an AI-native operating system he made for his own firm before opening it to other firms.
Episode 461 is mostly about plumbing: what a firm can pull out of its own systems, and how much that shapes what AI can do with it. We also got into how TORKLAW breaks into markets the billboard names already own.
Your biggest problem sits between departments. Jim's breakthrough came from looking at the seams instead of inside any one function. Intake, case management, phones, messaging, and marketing all worked. None of them talked, so his people became what he calls "the connective tissue," carrying data across by hand. It reached a boiling point three years ago, when TORKLAW had staff whose whole job was building reports, and every report landed a month out of date.
Ask how you get your data out. Jim wrote LawWorks the way he did because he'd been on the wrong end of the answer: Your data is your data, and you can walk whenever you want. That's the question to put to any system a firm evaluates now. Can you pull the structured fields out cleanly, and does the vendor share them out or hold them in?
AI output rises and falls on context. "Context is the thing that is shaping the quality of the output that you receive," Jim says. Storing the data isn't the hard part. TORKLAW's files lived in Box and its cases in CasePeer, and moving both by hand into a document generator meant the tool did its work with pieces missing. He wants a firm to push clean records out to a tool like EvenUp or Eve, then pull the finished work right back.
Precision over name recognition. Entering a market held by the billboard names, Jim goes narrow: paid search, organic search, and word-of-mouth referrals. He'll pay more per touchpoint as long as the acquisition price still works. What he won't assume is that any of it transfers. Something working in Illinois can fall flat in California, so he tests for the language that lands there and makes sure intake answers the phone on the same message.
The bottleneck moves. Solve the pinch in one place and it shows up somewhere else, Jim says, which is why he hires the Moneyball way instead of chasing A players. He'll take the intelligent problem-solver who's light on experience, pair them with the veteran who stalls before the finish, and coach the rest. He loves the puzzle and hates the process, and he still calls building a team fun.
"You end up with Zapier spaghetti and you end up with spreadsheets all over the place to fill in gaps and you have copy and paste errors and sync issues." — Jim Andresen, TORKLAW
The takeaway for PI firms: Find the report somebody at your firm rebuilds by hand every month. That's the seam where your systems stop talking, and it's the one your next software purchase should have to close.
Here's our full conversation:

🤖 AI SEARCH TIP OF THE WEEK
The pages AI cites for legal questions usually aren't law firm websites. They're the legal Q&A platforms. Crawlers hit them constantly, and models cite them heavily and trust the answers. Attorneys answering questions there publish inside a source the models already trust, on domains far stronger than their own. One PI firm made that a weekly habit as part of a broader AI search push and watched its ChatGPT visibility climb 992%.
The action this week: Pick your strongest practice area and answer two or three real questions on Avvo, in the words an injured person would use rather than the ones you'd put in a brief. Sign each one with your name and your firm. Do it every week and the answers pile up inside a source AI already reads. Earning that much trust on your own domain costs far more.

🛠️ TOOL OF THE WEEK
DingDuff Puts Millions of Cases and Statutes Inside Claude for Free
Two practicing lawyers built a free Claude connector for legal research. They say it beats the Westlaw AI their own firms pay for.
DingDuff links your existing Claude account to millions of court opinions. Its own database adds statutes, regulations, and court rules for the federal system, all 50 states, and D.C.
Case law comes from CourtListener, filings from PACER. Nothing new to learn. You ask Claude a legal question inside Claude, and it answers from the primary sources with citations you can open and check.
Kyle Dingman, a plaintiff-side mass-torts lawyer in Austin, built it with Stephanie Duff-O'Bryan, a commercial litigator and former Consumer Financial Protection Bureau enforcement attorney. Neither is a professional coder. The first version ran on a $35 Raspberry Pi in Dingman's closet.
Dingman told LawSites that an ordinary Claude account with DingDuff produced research memos matching his firm's legal AI. On their site, the founders say both stopped using their Westlaw AI accounts.
The statutes separate it from the free alternatives. Give a model only court opinions and it infers what a statute says from the cases quoting it, which breaks when the court skipped an exception or read a version since amended. Texas closed off reverse veil-piercing against LLCs in 2023, after a man dodged a $385,000 personal injury judgment by parking his assets in his own LLC. Three of the four systems DingDuff tested missed that statute and cited dead case law.
The citation checker runs on code. After you draft, it downloads every cited case and statute and pairs each claim with the passage supporting it. Because code does the matching rather than the model's memory, a misremembered quotation cannot pass as verified. In a demonstration for Bob Ambrogi of LawSites, it tracked 79 citations in one memo and flagged every authority it could not retrieve.
It's free, and the founders took no investors. You need your own Claude plan. The connector and its four skills run on a tip jar. The terms reserve the right to add fees on 30 days' notice.
Lawyers only, on purpose. Signup requires confirming you're a licensed attorney, because the founders built almost no guardrails in. They say they've never seen it invent a case, but it does misread them: "The kinds of mistakes it makes are the kinds of mistakes associates make,” Dingman told the LawSites blog.
Dingman calls DingDuff "just a pipe from the law to Claude." It hands the model the sources and stays out of the analysis, which makes the lawyer the only guardrail.
Read the benchmark with their caveat attached. Claude Fable 5 with DingDuff scored 11 of 11 on legal accuracy, against 9 of 11 for Westlaw CoCounsel and 4 of 11 for Lexis Protégé. Its citations came back 131 for 131.
But DingDuff wrote the assignment, built the answer key, and graded the memos. It's one prompt in one practice area.
Duff-O'Bryan still won't file anything without reading the cases herself and running a separate citator check.
🔗 DingDuff →
Disclaimer: Personal Injury Mastermind takes all reasonable steps to ensure accuracy in the materials we share, including articles, newsletters, and reports. These materials are intended for general informational purposes only and do not constitute legal advice. They may not reflect the most current laws or regulations. Always consult a qualified attorney for advice on a specific legal matter.

Thanks for reading. Quick ask…if you know someone who’d benefit from this content, please forward this to them. I’ll be back next week. - Chris
Received this newsletter from someone else? Subscribe below. Questions or want to sponsor this newsletter? Contact us at [email protected].


