Map the real process before you automate anything. Run Kravchenko's test yourself. Ask the owner how a job gets done, then ask the person doing it. The answers split because most firms run on habits, not enforced SOPs, and the process looks different from one desk to the next. Diagram the real workflow, get everyone who touches it to agree on it, then decide where a tool fits. Automate a process nobody has written down and you just automate the chaos.

Below 25 people, technology is not your bottleneck. This is Kravchenko's least popular point and his most useful. A 30% efficiency gain on a three-person team does not add up to one full-time employee. The same 30% across a hundred-person team is thirty people. Under that size, the constraint is staffing, training, and customer service, not software. Fix the human system first.

The expensive tool is often the cheap one used well. Kravchenko spent three months testing legal-AI vendors built for personal injury. For routine cases, he found the specialized product running a few hundred dollars per case was sometimes better than a general model, sometimes worse, and rarely worth the premium over a standard subscription plus a few extra minutes of work. Every one of them still needs a human to check the output, so any vendor promising "hallucination-free" cannot possibly live up to that claim. Put the money into your team's skill with the tools they already have.

Watch the one workflow that decides your margin: medical records. A serious case can involve thousands of pages of medical records, and Kravchenko found that past roughly 300 to 400 pages in a single chat, the model starts making errors of omission. It does not invent a treatment. It quietly drops one. His fix is a funnel: Split the records by date, summarize each batch in its own chat, then draft the chronology or demand from the twenty clean pages of summaries. You end up with a chronology you can check instead of a summary you just have to take on faith.

Map every page to intent, not volume. One keyword, one page. Each case-type long-tail gets its own practice sub-page. The deadline terms get a state-specific page. Stop pointing five high-intent terms at one bloated service page where they fight each other.

Build the page everyone else skips. Answer "how long do I have to file" for every state you practice in. Deadline up top, call-now path right under it. The state-specific version is high-intent and far less contested than the generic term the big firms chase.

Run your gap analysis on the specifics, not the generics. Pull the case-type and deadline keywords your top local competitors rank for and you don't. We do this for Rankings clients constantly, and the long-tails rank in 90 days where the generic term takes years and a larger budget.

Grade your SEO by signed cases per page, not traffic. A page that draws 50 visitors and signs three catastrophic cases beats the guide that draws 5,000 researchers and signs none. Traffic is a vanity number. Signed cases are the one that pays.

Private equity is moving into professional services, and PI is on the list. Investors want steady revenue and businesses where AI can lift margins, Reuters reported, and law firms fit the bill. A billion-dollar check aimed at professionalizing and scaling the country's biggest PI firm makes it a tougher competitor in every market it touches.

A Morgan & Morgan deal would set the template. Smaller firms already took outside money this way. We covered private equity's move into PI firms earlier this year. Morgan & Morgan would test the structure at the top of the market, and a deal at the largest firm in the country would normalize it for everyone below.

This is firm ownership, not case funding. Keep them separate. In our last issue we covered the federal tax bill aimed at litigation funders, the capital behind your toughest cases. This is the other capital story: private equity buying into who owns the firm, not who bankrolls the docket. Different money, different risk, different rules.

Per-case lead deals in Colorado become illegal August 12. The ban covers the pricing models the legal lead industry runs on: per lead, per case, and subscription. A Colorado firm with a vendor contract on any of those terms has about eight weeks from this edition to replace that case flow with marketing under its own name.

Other lawyers enforce it. The $10,000-per-violation private right of action belongs to competitors and consumers alike, and Cheney has already announced day-one lawsuits. A vendor still selling leads into Colorado on August 13 faces $10,000 claims from every firm and consumer it touches.

Who the ban favors is the live fight. Rhoden's argument is the one a firm owner feels: the channels that stay legal, TV, billboards, SEO, brand campaigns, reward a seven-figure budget, and per-case lead buying was the entry point for firms without one. Whether the law protects injured consumers or entrenches the biggest advertisers is the question the plaintiffs' bar is now arguing in public.

PI Firm Hits Uber and FedEx With Counterclaims Over RICO Suit: Philadelphia lawyer Marc Simon and his firm Simon & Simon filed counterclaims accusing Uber and FedEx of abusing the federal racketeering law to drive him out of business and "chill other personal injury firms." Simon argues the companies settled the same injury cases in state court while branding them fraud in their federal RICO suit, which he calls evidence the action is "objectively and subjectively baseless." Uber and FedEx sued Simon last year, alleging he steered clients to a "network of corrupt medical providers" who inflated claims. A judge denied Simon's motion to dismiss last month. Uber said the counterclaims "lack merit" and it will move to dismiss. The companies have filed similar suits against firms in Los Angeles, Miami, and New York, Reuters reported.

Rival PI Firm Hits Isaacs & Isaacs With Federal Antitrust Suit: Alex R. White PLLC and attorney Nicholas Alexiou sued Isaacs & Isaacs in federal court in Louisville, alleging the firm's employment agreements violate federal antitrust law by claiming 70 percent of fees on clients who follow departing lawyers, imposing an $8,000-per-client penalty, and restricting client contact. The Isaacs firm has sued at least five other associate-level lawyers and their new employers over the past 18 months, the lawsuit said. The plaintiffs seek treble damages and a court order barring enforcement of the agreements. Isaacs & Isaacs, which sued Alexiou last year over alleged client poaching, called the new filing "a desperate ploy to misdirect away from his unprofessional conduct." Reuters reports the case is among the first antitrust suits over law firm employment terms.

Kirkland & Ellis Commits $500 Million to Build Its Own AI Platform: Kirkland & Ellis, the highest-grossing law firm in the world, will spend about $100 million this year and "hundreds of millions more" over four years building proprietary AI rather than relying on tools its competitors can also buy, chair Jon Ballis told the Financial Times. Kirkland is building the platform on the work of 250 of its lawyers, and the outside companies helping develop it cannot sell the technology to anyone else. Kirkland is funding the investment from its own revenue, a record $10.6 billion in 2025, more than any firm globally. Its first launched product, built with Palantir, targets private equity fund formation across the firm's 1,000-plus investment-funds lawyers.

Jury Awards $13 Million to Nichelle Nichols's Estate in Wrongful Death Case: A New Mexico jury awarded $13 million to the estate of Star Trek actress Nichelle Nichols in a wrongful death lawsuit against Gila Regional Medical Center in Silver City, which treated Nichols in 2022. The verdict sheet assigned 40 percent of the liability to the hospital and 60 percent to one of her doctors. Nichols, who had a heart condition, allegedly did not receive proper treatment during an emergency-room visit and died hours after a transfer to an assisted-living facility, KOAT reported. She was 89.

Abbott Must Face Class Action Over PediaSure Height-Growth Claims: A Manhattan federal judge ruled Abbott Laboratories must defend a proposed class action claiming it misled consumers into believing PediaSure Grow & Gain drinks were "clinically proven" to help children grow taller. In a 75-page decision, U.S. District Judge Paul Engelmayer said the label's cartoon giraffe and ruler-like marks could support reading "grow" as height growth, and jurors could find Abbott's commercials reinforced that message. A Bronx grandmother filed the suit in 2023, saying her grandson stayed short for his age but became "so overweight" that she stopped buying the drinks. Abbott called the labeling "appropriate" and said the evidence will show the allegations are unfounded.

JustPoint is an AI company that became a law firm. The venture-backed startup, which lists more than $105 million in funding, uses AI and medical data to surface toxic exposures in drugs and products, the kind of early warning that starts a mass tort. It opened an Arizona law firm to represent those clients directly. Arizona is one of the few states that lets non-lawyers own a firm.

Leathers is building the tools to find the cases first. The bootstrapped firm runs Velocity Justice, whose platforms score drug-safety signals and aggregate litigation data to, in its words, "find the case first" and "market to plaintiffs before anyone else." Bellwether scoring, settlement-waterfall simulations, automated consolidation motions.

An attorney connects with every potential client. When a good lead comes in, her intake team posts it to a Slack thread and whichever attorney is free jumps on the call. The attorneys are partners who take a share of the cases they sign, so the incentive and the firm's values point the same way. Carol-Lynn's read is direct: When a client builds a relationship with the lawyer from the first call, the average fee per case goes up.

The intake call turns on one question. What matters most to this client, right now? Sometimes it is meeting the lawyer. Sometimes it is medical treatment. Sometimes it is a rental car so the kids get to school next week. Her team identifies that one thing on the call and solves it immediately. She credits that habit, more than any script, for the conversion rate.

She replaced case managers with phase specialists. In 2022 the firm scrapped the traditional one-manager-per-case model and reorganized around case phase. A liability investigator sizes up the available coverage, a separate teammate manages treatment, then the file moves to a demand writer, a negotiation team, a settlement coordinator, and litigation if it gets there. More hands per case, each an expert at one stage. She runs it on Filevine, configured to add the right teammate as the case changes phase, which keeps the handoffs from turning into silos.

Trial credibility raises the settlement. Carol-Lynn hired former insurance-defense lawyers who know the carrier playbook, and the firm staffs board-certified civil trial attorneys. Insurance carriers price that in. When the other side knows a firm will take the case to verdict and hold out for full value, the offers come up before trial.

Decide who is qualified to take it.

Solve the client's single biggest worry on that call.

It captures the case-value facts a message-taker misses. On a single call, ClaireAI confirms the incident facts, police report, fault, and treatment status, plus every insurance layer, PIP, BI, MedPay, and the UIM/UM coverage it says it is calibrated to pull even when the caller only volunteers "I have insurance." Underinsured and uninsured motorist limits are the case-value drivers a generic receptionist never thinks to ask about. On a pre-cleared grade-A lead, ClaireAI e-sends a firm-branded DocuSign retainer, often signed before morning, the company says.

It runs legal-aware triage on statute and conflicts. ClaireAI claims jurisdiction-by-jurisdiction statute-of-limitations windows, and routes any call within 90 days of a deadline to a partner instead of auto-rejecting it. Before booking, it collects the caller, opposing party, and insurance carriers and runs a fuzzy-match conflict check against your CRM, declining to schedule when it flags one, per ABA Model Rules 1.6 and 1.18.

It grades and routes every lead, then writes to your CRM. ClaireAI scores each call A through D on liability, treatment, policy limits, statute timing, and conflicts: A goes to a partner within five minutes, B gets a consult-booking text, C enters nurture, D routes to a referral partner. It writes the contact, matter, brief, and recording into CASEpeer, Filevine, or Litify during the call, with no Zapier middleware.

It carries TV and mass-tort spikes that break a phone desk. Every caller reaches Claire at once, with no busy signal or queue, whether the firm gets three calls an hour or three hundred in the ten minutes after a spot airs. That concurrency is the structural difference from a one-call-at-a-time receptionist or a shift-staffed answering service.