The Pre-Intake Problem in the News – July 28, 2026

We keep hearing stories of injured people turning to AI before they ever reach out to a law firm. This week the story comes from the other side – the attorneys who defend insurers.

Two insurance defense attorneys published a warning to insurers last week about people who sue without a lawyer.

Craig Compoli and Brian Griffin defend insurers at O’Toole Scrivo, a New Jersey trial firm. Their article in the National Law Review looks at what happens when self-represented (pro se) litigants show up with AI-drafted pleadings.

Their worry, in their own words:

“The concern is not whether an AI-generated filing will ultimately fail, but how much defense work is required to test, narrow, or defeat it.”

Winning still costs them more than it used to.

The numbers behind the article come from a University of Miami study of about 2.8 million federal court filings, by Or Cohen-Sasson of the Miami Law & AI Lab.

Roughly 1 in 6 federal civil cases is now filed by a person without a lawyer. Before generative AI it was about 1 in 9 (11.33% to 16.94%, per the study).

Last issue we cited a different study that found the same increase. Two research teams, working separately, measured the same thing.

The AI-drafted complaints looked more polished, cited more law, and still got dismissed earlier than other pro se cases. The insurers paid to fight them anyway.

The problems in the article are real. The authors walk through six federal decisions dealing with AI-generated errors in pro se filings, made-up citations among them, and courts are starting to require people to disclose when AI helped write a filing.

So who are these people?

Gerry Spence answered that in his book Win Your Case:

“In a single week I will receive scores of requests for representation — mostly from ordinary citizens who have just cases with small money damages or complex cases that are too time consuming and costly to bring.”

One of the most famous trial lawyers in America, turning away scores of people a week because their cases were too small or too expensive to take. He wrote that before AI existed.

These people didn’t appear when ChatGPT was created. They were always there, with no way to bring their cases. Now they have a tool, and the court data is starting to show how many of them there were.

A comment on one of our earlier posts said AI has no place in law.

Compoli and Griffin’s own article says otherwise:

“For people who cannot afford counsel, AI may seem like a low-cost substitute for legal representation.”

For the person whose case is too small for any firm to take, AI is the only help they can afford.

Spence again from the same page of his book: “That there is ‘liberty and justice’ for all is the great myth of America.”

In personal injury, most of this never becomes a court filing. A person with a smaller or uncertain case asks AI whether they have a claim, gets an answer, and handles the adjuster on their own, or lets it go. They never call a firm.

The insurance industry is reading about these people in its trade press now. Carriers are being told to spot AI-driven filings early and budget for them.

Insurance companies are adjusting to people who start with AI instead of a lawyer. Most personal injury firms are still waiting for the phone to ring.

What happens in that window, from injury to intake, is what we call The Pre-Intake Problem™ – and it has been largely unserved until now.

If you run a personal injury firm and want to know what AI is telling injured people in your market – who it recommends, who it leaves out, and what that means for the cases you sign – request a Pre-Intake Revenue Strategy Discussion. Thirty minutes on your market and your revenue, no pitch. Request one at lawshift.ai.