👋 Good morning. Chris Dreyer here. ChatGPT referral traffic grew 206% year over year, and Semrush data shows those leads convert at 4.4x the rate of organic search. The person who clicks a ChatGPT recommendation already described their situation, got answers, and narrowed their options before they reached your site. This week I break down the data and the content architecture that gets your firm into those recommendations.
Also this week: Google replaced call duration with an AI model that scores what your callers actually say. If your firm runs call extensions, the way Google counts your leads just changed. And I pulled apart Morgan & Morgan's Facebook ad portfolio. They run roughly 1,100 active ads, and almost none promote car accidents. What they are promoting tells you where the largest PI firm sees the money going.
Plus: Stanford researchers found that only 22 of 500 PI firm websites disclose a fee percentage, and Uber is using that opacity to push a fee cap in California. 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
Lock in your ticket today for the best prices
💡ONE BIG IDEA
ChatGPT Bolsters Case for Rebuilding Your Content Around AI Search
More injured people are finding lawyers through ChatGPT. Semrush data shows those referrals convert at 4.4x the rate of organic search.
The reason is simple. The person who reaches your site through ChatGPT already described their situation, got answers, and narrowed their options before they clicked. They arrive self-qualified. No other channel delivers a lead that far along in the decision before the first conversation with your intake team.
Semrush analyzed more than 1 billion lines of U.S. clickstream data across a 200-million-user panel over 17 months. ChatGPT referral traffic grew 206% year over year. ChatGPT now sends visitors to more than 170,000 unique domains every month.
An accident victim types, "I was rear-ended on I-35 in Dallas, the other driver was texting, I have back pain and missed two weeks of work, what should I do?" ChatGPT walks them through their options, explains what a personal injury attorney handles, and recommends specific firms.
By the time that person picks up a phone, they already know what kind of case they have, what kind of help they need, and which firm ChatGPT told them to call.
The conversion lift holds across sectors. Visibility Labs tracked 94 brands over 12 months and found ChatGPT traffic converted 31% better than non-branded organic search. One B2B professional services firm measured a 15.9% conversion rate from ChatGPT referrals versus 1.76% from Google organic, a 9x lift. The pattern is consistent: AI-referred visitors act at higher rates because they arrive with intent already formed.
65% of ChatGPT queries never trigger a web search. Those two-thirds pull answers from training data alone. If your firm's content, reviews, and case results do not exist in that training data, ChatGPT has no reason to mention you.
In early April, I wrote about share of model: the percentage of AI queries in your market that surface your firm. That piece focused on what AI agents look for when they decide which firm to recommend. This week I want to focus on the content architecture that feeds those recommendations.
A good framework for this is the hub-and-spoke model that Animalz, a content strategy consultancy, has written about for years. A hub page is not a blog post. It is a table of contents for a topic, a page that organizes and links to every spoke page covering a subtopic beneath it. Each spoke targets one specific question a potential client would ask.
Build a hub page for each core practice area. If your firm handles car accidents in Dallas, the hub links to every spoke you have: what to do after a rear-end collision, how fault works in Texas, what compensation looks like for back injuries. I have seen firms rank for dozens of long-tail queries they never targeted directly, simply because the hub gave AI models a reason to treat the site as an authority.
Each spoke page should answer one question directly. "Can I sue if the other driver was texting?" "How much is a rear-end collision case worth in Texas?" Answer in the first paragraph. ChatGPT pulls from sources that give direct, authoritative answers, not from pages that bury the answer under 1,500 words of background.
Make your content machine-readable. Implement an llms.txt file that maps your practice areas, geography, and credentials in a format AI agents can parse. I covered llms.txt in the April 6 issue; the AI Search Tip below walks through exactly how to build one.
Audit your visibility now. Open ChatGPT and ask it to recommend a personal injury lawyer in your primary market. Record what it says. That is your baseline.
OpenAI is already treating this channel as a performance platform. In February, we covered OpenAI rolling out ads inside ChatGPT. By April the platform crossed $100 million in annualized ad revenue.
Now Digiday reports that OpenAI is building a conversion pixel, the same tracking infrastructure Meta and Google use, to let advertisers measure whether ChatGPT ads produce completed actions on their sites. The pixel is already live for pilot advertisers.
That tells you where this is headed. The paid layer is coming. The firms that show up in ChatGPT recommendations organically today have a head start before paid placement crowds the results.
At Rankings.io, we are structuring client content so AI models can find it, read it, and recommend it. If your firm is not in these results yet, this is the issue to forward to whoever owns your content strategy.
🔗 Semrush · Visibility Labs · Digiday →
♟️STEAL THIS PLAYBOOK
Morgan & Morgan Has Roughly 1,100 Active Facebook Ads. Here Is What the Current Portfolio Shows.
I looked at Morgan & Morgan's active Facebook ad portfolio this week using Meta's Ad Library.
The current portfolio is almost entirely mass tort. Dexcom G7 recalls, 9/11 exposure claims, toxic cabin air lawsuits, Depo-Provera, Ozempic vision loss, DoorDash data privacy, GM transmission defects, processed foods linked to childhood diabetes. Car accidents, slip-and-falls, and premises liability barely appear.
That is a snapshot, of course. Campaigns are bound to change. But right now, the largest PI firm in America uses Facebook to fill a mass-tort funnel, not a general PI one.
Seventy percent of the ads use designed graphics with text overlay, not video. Roughly 770 of the active ads are image-with-text. Fewer than 320 are video. Fewer than 10 are static images. The highest-impression ads are bold text on dark backgrounds. Cheap graphics at volume, not polished production.
One creative concept has up to 23 ad variations live at once. A single work-injury ad carried 23 active versions. A DoorDash privacy ad carried nine. A disability appeal carried six. Same core copy, different tweaks, different audience segments. Meta's algorithm picks the winners.
Quiz funnels pre-qualify every lead before it reaches intake. Ad after ad sends traffic to forthepeople.com eligibility quizzes. The quiz filters contacts before anyone picks up the phone. Low-value leads never get through.
An anti-runner campaign doubles as brand defense. A cluster of ads asks "Contacted by a Lawyer at the Scene or Hospital?" and warns viewers about illegal solicitation. The ads protect the brand against barratry claims and position the firm as the ethical alternative where runners erode client trust.
Retarget ads confirm a layered funnel. At least one active ad reads "Still Interested? Finish Signing Up Today," aimed at users who abandoned a quiz. Awareness ad, quiz, retarget. Not a one-touch system.
Do you know what your competitors run in your market? Try this.
Meta's Ad Library is one of the few competitive intel tools that gives you this level of detail for free. Every active ad, every creative format, every launch date, all public. If you are a newer firm or have never done this exercise, start here:
Go to facebook.com/ads/library. Search the competitor's page name, not a keyword. Select the advertiser from the dropdown to filter to that single page.
Note the total active ad count. Sort by total impressions to surface the top performers.
Filter by media type (video, image, image-with-text) to map the creative format mix.
Scroll the top-impression ads and categorize by case type. Track which practice areas the competitor funds on paid social.
Check launch dates. Ads active for months are proven performers the competitor keeps funded.
Count versions per creative. High counts mean active tests. Low counts mean set-and-forget.
Read the copy for patterns: CTA language, social proof claims, conversion path (phone, form, quiz).
Run this across your top five local competitors and the national players in your market. Within an hour, you will know which case types they prioritize, what creative formats they use, how aggressively they test, and where the gaps are in your own strategy.
The data updates in real time. It costs nothing. And your competitors have no idea you looked.
🔗 Meta Ad Library →
📰 TOP OF THE NEWS
Google Ads Replaces Call Duration With AI to Decide Which Leads Count as Conversions
Google Ads no longer counts call duration to decide which leads are conversions. A new AI-Qualified Call Conversions feature, announced this week, replaces duration-based tracking with an AI model that scores the actual content of each recorded conversation.
The old system counted any call that exceeded a set time threshold (often 60 seconds) as a conversion, whether the caller asked about legal representation or dialed the wrong number.
The new model evaluates what the caller said: questions about specific services, requests to schedule a consultation, direct statements about intent to hire. Robocalls, spam, and misdials no longer count.
Smart Bidding now optimizes toward AI-qualified leads, not call length. Google feeds the new classification data directly into its automated bid algorithms. PI firms that bid on call extensions or call-only campaigns will see Google reallocate spend toward conversations the AI scores as high-intent. That changes cost-per-lead math for every firm that buys paid search phone inventory.
Firms that turned off call recording lose the upgrade. Google enabled recording by default for most advertisers, but accounts classified as healthcare or financial services still have it off. PI firms that operate under healthcare-adjacent compliance settings should verify their account status now. Without the audio, the AI cannot evaluate conversation quality, and Smart Bidding falls back to the old duration metric.
Google now produces call summaries with intent tags. The Call Details report surfaces labels like "#HighIntent" and "#ConsultationScheduled" alongside a plain-language summary of each conversation. Intake managers can audit whether Google's AI classification matches their own lead-quality standards and adjust duration thresholds if valid leads fall through.
The feature currently applies only to conversations where both numbers are in the United States or Canada.
🔗 Google Ads Help →
The Contingency Fee Model Is Worth Defending
🤬 Uber has some nerve. The ride-share company filed a California ballot initiative to cap PI contingency fees at 25%. People are fighting back.
A Stanford Law School working paper published in the Georgetown Law Journal opposes the cap. Researchers Nora Freeman Engstrom and Brianne Holland-Stergar argued in the Sacramento Bee that it would lock injured plaintiffs out of the courthouse.
But the paper also exposed something the industry cannot ignore: only 22 of 500 PI firm websites disclose a specific contingency fee percentage.
Almost none explain why it is worth defending. That’s a messaging problem.
Before I suggest a fix, let’s first look at what the paper revealed.
The fee market shows zero meaningful price variation. Engstrom and Holland-Stergar reviewed PI firm websites across all 50 states. Every firm that listed a rate quoted approximately one-third of the recovery. Only one firm in the entire 500-site sample advertised fees lower than its competitors. Three prior empirical studies that date to 1999 reached the same conclusion across case types, experience levels, and geographic markets.
The researchers compare PI's fee structure to real estate before the NAR antitrust settlement. Both markets feature sticky fees, aggressive non-price competition, first-time consumers who never shop, and almost no public data about what providers charge. The $418 million NAR settlement in 2024 forced brokers to disclose commissions and acknowledge that rates are negotiable. Engstrom and Holland-Stergar argue PI faces the same structural conditions and the same vulnerability to external disruption.
The paper recommends disclosure, not caps. Fee caps backfire. A 2009 RAND study of nearly 3,000 PI attorneys found that legislated limits reduce the stated willingness to accept cases. California's existing sliding-scale restriction produced a 46% reduction in attorney fees and made lawyers more selective about which clients they represent.
The alternative: mandatory "closing statements." The authors propose post-case filings that disclose gross recovery, attorney fees, expenses, and net to client for every resolved PI matter. New York has required similar filings for decades. Closing statements would arm prospective clients with benchmark data and open price competition in a market that has never had it.
Here's how we win: Publish content across text, video, and audio that explains exactly what the contingency fee pays for. Investigation. Expert witnesses. Court costs. Years of work before a dollar comes in. Name the risk we absorb so the client does not have to.
Then put that language in your intake materials, your advertising, and your community outreach. The firms that make the public case for why this model exists take the Uber argument off the table. The firms that stay silent hand the narrative to a company that will not tell it fairly.
🚀 QUICK HITS
Legora Taps Jude Law for Global Legal AI Campaign: Sweden-based Legora, valued at $5.5 billion after a $550 million raise last month, launched a celebrity ad campaign with Jude Law across London, New York, Los Angeles, and other cities. Swedish agency NoA Åkestam Holst created the spots, which lean on Law's name as a deliberate pun and position the legal AI platform as a long-term partner for law firms. Oscar-winner Hoyte van Hoytema shot the hero film and three supporting spots through Stink Films.
Florida Grand Jury Indicts Surgeon Who Removed Wrong Organ: A Walton County grand jury indicted Dr. Thomas Shaknovsky, 44, on a second-degree manslaughter charge in the death of William Bryan, 70, of Muscle Shoals, Alabama. Bryan visited Ascension Sacred Heart Emerald Coast Hospital in Miramar Beach in August 2024 with spleen pain. Shaknovsky removed his liver instead. Colleagues raised concerns about the surgeon's skill level before the procedure. Florida's Health Department suspended Shaknovsky's license in September 2024. He faces up to 15 years in prison. Arraignment is set for May 19.
Family of Kansas Jail Inmate Files Federal Wrongful Death Suit: Relatives of Charles Adair filed a federal lawsuit that names the Wyandotte County sheriff, the unified county government, and deputy Richard Fatherley, who faces second-degree murder charges in Adair's death. Investigators determined Fatherley shoved his knee into the handcuffed Adair's back for one minute and 26 seconds. Adair had been arrested on misdemeanor traffic warrants. Attorney Ben Crump, who represents the family, called for the public release of video footage the sheriff's office has declined to provide.
First Federal Conviction Under Intimate Deepfake Law: James Strahler II, 37, pleaded guilty to cyberstalking, production of child sexual abuse material, and publication of digital forgeries under the Take It Down Act, which President Trump signed in May 2025. Strahler used AI to generate nonconsensual images and videos of both adult and minor victims, posted over 700 images to a child sexual abuse website, and sent AI-generated material to victims' co-workers. The law requires online platforms to remove nonconsensual material within 48 hours of a report.
Uber Faces Second Federal Trial Over Driver Sexual Assault Claims: Uber returned to federal court in Charlotte, North Carolina, for a second bellwether trial after a jury in Arizona awarded $8.5 million to a plaintiff in the first case to reach a federal verdict. The plaintiff alleges an Uber driver grabbed her inner thigh and made suggestive comments when she reached her destination in Raleigh in March 2019. Uber argues it operates as a software company with no legal duty to protect passengers. U.S. District Judge Charles Breyer oversees more than 3,300 consolidated lawsuits in the same litigation. The bellwether verdicts will help determine claim values for any potential settlement.
💯 NUMBER TO NOTE
AI summaries in search results cut click-through rates to ranked websites nearly in half, from 15% to 8%. That is the finding from a March 2026 Harvard Business Review article by Graham Kenny and Ganna Pogrebna. The pages your firm spent years to build now lose almost half their clicks when Google answers the question before the searcher reaches your site.
The authors document two forces that reduce friction for consumers and increase it for businesses: LLMs like ChatGPT replace search engines as the first place consumers go for answers, and Google's own AI Overviews absorb the clicks that used to flow to ranked pages. For some publishers, click-through rates have fallen by as much as 89%.
The research trip that used to take 15 to 20 website visits now takes one AI response. The HBR article profiles a health insurance provider that tracked how consumers research coverage options. Questions that once sent visitors to educational pages, policy explainers, and comparison tools now get answered in a single ChatGPT exchange. No link back. No brand exposure. The repeated visits that built brand recognition vanish with them.
Your content now serves two audiences: the reader and the model. The article makes this case directly. In a market where AI decides which firms to cite, clear structure, named experts, specific credentials, and consistent descriptions across every surface where your firm appears all determine whether the model includes you in its answer.
This connects to the case I made in One Big Idea this week. The 8% should change how PI firms evaluate their web traffic. The pages that rank on Google still matter, but the clicks they generate shrink every quarter as AI answers the question before the searcher reaches your site. The firms that build content AI models can find, read, and recommend will capture the demand that ranked pages no longer deliver.
🎙️ FROM THE POD
Eric Elliott on the Case for Investing Before You Advertise
Eric M. Elliott has spent 20 years in media and legal marketing. He believes the firms that dominate their markets all do the same thing: They get the sequence right before they spend a dollar.
Eric runs VIP Marketing and Craft Creative. On Episode 399, we talked about the three fundamentals he requires before any campaign goes live: budget, media, and message. In that order.
Get the budget right first. If your firm cannot afford to compete in a channel, that channel is the wrong choice. Then pick the media. Then craft the message. Eric treats those three as gates. All three clear before a dollar moves.
Intake is a marketing problem, not an operations afterthought. Eric put it plainly: You can send 600 leads to a practice, and if the front desk answers late or skips the empathy, the agency takes the blame. Agencies do not sign cases. Firms do. Any practice that burns through ad spend without a hard look at its own close process measures the wrong variable.
Meta is where you learn what your market will buy. You can test every case type on Meta: nursing home, dog bites, auto, each with its own creative and funnel, for a fraction of what broadcast costs. Adjust the message in hours, not weeks. Eric's phrase: You can A-through-Z test on Meta. No firm should scale a channel it has not pressure-tested first.
Strong firms build referral loops, not just ad funnels. Eric credits the flywheel concept to others, but his application to PI is specific. Good marketing attracts the client. The firm delivers a fair outcome. That client tells everyone at church, at work, at the neighborhood barbecue. The referral feeds the next case. He sees the same gap everywhere he consults: A media budget exists, but no referral budget and no line item for the community that sustains the practice.
Community deposits earn what paid media cannot. Eric shared the story of Joy Law Firm in South Carolina. More than 12 years ago, Joy's attorneys stepped in to defend a high school football team that was denied a shot at the state championship. They took the case to the South Carolina High School League. No money changed hands. The entire town remembered. Eric calls this a deposit: You put value into the community before you ask for anything back.
I asked Eric what question every firm should answer before the next budget cycle. His answer was simple: Are you making deposits, or just withdrawals?
"You are asking for withdrawals from the community, but you're not making any deposits."
Here's the full conversation with Eric:
🎧 Personal Injury Mastermind: Episode 399 →
🤖 AI SEARCH TIP OF THE WEEK
Build your AI search strategy on infrastructure you own, not platforms you rent. Chris Mele put it directly in Fast Company: "GEO is not a new strategy. It's a new lease on the same building." McKinsey's own research shows a brand's owned content accounts for 5 to 10 percent of what AI search actually references. The rest comes from affiliates, publishers, and sources the brand does not control. Optimizing for AI Overviews means polishing a minority stake while the platform keeps the audience.
The action this week: Audit your content strategy for the landlord problem. If every channel you invest in (SEO, GEO, AIO) depends on a platform you do not control, you are renting. The alternative is owned infrastructure: your website, your podcast, your email list. Build something worth finding, and AI search visibility follows. The firms that survive every algorithm change built the relationship before the landlord renovated.
🛠️ TOOL OF THE WEEK
Claims Concierge Takes Property Damage off Your Desk So Your Team Works Injury Cases
Every auto accident PI case comes with a property damage file attached to it. Diminished value claims, total loss disputes, rental car coordination, repair shop oversight, insurance adjuster calls. None of it generates contingency fees. All of it eats paralegal time.
Claims Concierge, launched in 2024 out of Atlanta, takes the entire property damage workstream off your plate. Your firm refers the property damage component. Claims Concierge handles insurance coordination, diminished value recovery, total loss advocacy, repair shop management, and rental car acquisition. No upfront fees. One transparent fee on the backend.
Claims Concierge takes over property damage operations for PI firms that want their paralegals focused on injury cases, not on adjuster calls over rental car receipts.
Diminished value recovery targets money clients rarely collect. A repaired vehicle with accident history loses $3,000 to $6,000 or more in resale value. Clients rarely file a diminished value claim because no one tells them it exists. Claims Concierge files the DV claim against the at-fault driver's insurer and recovers compensation the client did not expect.
Total loss advocacy challenges the insurer's first offer. Insurance companies apply standardized depreciation formulas that routinely undervalue vehicles. Claims Concierge reviews the valuation report, identifies mismatched comparables and inflated depreciation adjustments, then advocates directly with the carrier for a corrected figure.
Repair coordination and rental car acquisition remove two of the most common client complaints. Claims Concierge connects clients with a vetted repair shop network and arranges a comparable rental vehicle. Clients stay mobile during repairs. Those calls disappear from your intake staff's day.
The partnership model turns a cost center into referral revenue. Your firm refers property damage work it already has. Claims Concierge handles every step. You earn referral fees on cases you could not monetize before and eliminate the staff hours you devoted to them. No new headcount. No new software.
For firms that handle 50 or more auto accident cases a month, count the paralegal hours your team spends on property damage admin. That time has a dollar cost, and it produces zero contingency revenue. Claims Concierge converts it into a revenue line.
🔗 Claims Concierge →
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? Contact us at [email protected].