I have spent much of my career looking at insurance company claims manuals, training materials, management directives, performance measurements, and other documents that tell us how an insurance company really handles claims. The individual claim file is important, but sometimes it is only the last chapter of a much longer story. A recently filed motion in the Oklahoma State Farm hail litigation is a good example of why policyholder attorneys need to look upstream.

A filing in Hursh v. State Farm 1 seeks to compel non-party Accenture LLP to produce documents concerning consulting work it performed for State Farm. Before anybody gets too far ahead of the evidence, this is a plaintiffs’ motion. Many of the underlying exhibits are under seal, State Farm and Accenture dispute important aspects of the plaintiffs’ characterization, and there has been no judicial finding that either company engaged in improper claims conduct.

With that lawyerly disclaimer out of the way, there is some fascinating stuff in this motion.

The plaintiffs contend that State Farm paid Accenture more than $1 million for work involving State Farm’s wind and hail claims practices. According to the filing, an April 2020 Accenture presentation identified six Accenture employees working on the project, and subsequent materials involved claim type, jurisdiction, contractors, damage identification, hail size, and matching weather information with reported losses. That sounds like a lot more than hiring somebody to tell management whether everybody likes the coffee in the claims department.

The plaintiffs’ central contention is that Accenture helped State Farm build parts of the process and measures for handling these claims. The motion says Accenture developed claim-type analyses in which certain partial-roof outcomes were treated as normal while full-roof replacement was characterized as an anomaly. The motion also connects those analyses with a greater likelihood that a claim would close without payment. Those are the plaintiffs’ characterizations, and the documents themselves need to be examined before anyone treats those characterizations as established fact.

Still, it raises a very important question about claims management. Was State Farm trying to make individual claim decisions more accurate, or was management identifying a desired aggregate result and then developing claims procedures intended to produce that result?

There is nothing wrong with an insurance company studying its claims results. Insurers should look for fraud, waste, overpayments, inconsistent adjusting, mistakes, underpayments, and poor training. I would be more worried about an insurance company that never studied any of those things.

The problem comes if the analysis starts at the other end. Suppose somebody determines that the company is paying too much for roofs, calculates how much money could be saved by bringing that percentage down, and then creates procedures, approval requirements, benchmarks, and management measurements designed to reach the desired percentage. At that point, the obvious question is whether the facts of the individual claim are driving the result or whether the desired financial result is influencing the facts the adjuster is permitted to recognize. That is why the financial language described in this motion interests me so much.

The plaintiffs say Accenture quantified the financial impact of reducing full-roof replacements in the tens of millions of dollars. The motion also refers to an Accenture document that allegedly identified a $169 million Texas “opportunity.” Much of the context surrounding that figure is redacted, so nobody should pretend we know exactly what it means yet. But $169 million is not a number that slips unnoticed into the corner of a PowerPoint presentation.

This is where discovery becomes critical. What was the “opportunity”? How was it calculated? What assumptions went into it? Was the calculation based upon eliminating genuine overpayments, or was a reduction in indemnity itself being treated as a measure of success? Most importantly, did those financial calculations have anything to do with the claims procedures eventually imposed upon individual adjusters?

Those are fair questions. They are also questions that can’t be answered by interviewing the adjuster who climbed on one policyholder’s roof years later.

The management-approval allegations are equally interesting. Plaintiffs cite State Farm materials which they say show that after an initial Texas effort, wind and hail tactics were implemented more broadly in January 2021 and that team-manager approval was required for certain full-roof replacements involving damage potentially caused by smaller hail or lighter winds. Plaintiffs contend the Hursh claim was adjusted under those rules.

There is nothing inherently sinister about supervision. Claims managers are supposed to supervise adjusters. Expensive claims deserve careful review, and quality assurance is part of running a responsible insurance company.

But again, the purpose of the review matters. If the manager is asking, “Did we get this claim right?” that is good claims management. If the manager is asking, “Why are you replacing another roof when we are trying to reduce the percentage of full-roof replacements?” we have a very different issue. That is the type of distinction the underlying documents may reveal.

Another part of this filing should cause policyholder lawyers to rethink where they look for evidence. According to the motion, Accenture’s final materials contemplated measurement reports, claim prioritization, early-signal monitoring, root-cause analysis, and accountability regarding tactic usage and adoption. The plaintiffs also allege that an Accenture consultant used an assigned State Farm email account while helping organize work and mobilize State Farm personnel around the program.

Claims lawyers love claims manuals. We subpoena them, fight about them, and occasionally treat them as though Moses carried them down from the mountain.

Maybe we should spend a little more time asking for the PowerPoints.

I would want the dashboards, scorecards, executive presentations, performance measurements, adoption reports, financial analyses, internal emails, consultant workpapers, and whatever management was reviewing when somebody reported whether the program was working. Sometimes a polished claims manual tells you what an insurance company says should happen. A management dashboard may tell you what the company was actually trying to make happen.

The use of the phrase “industry standard” also deserves scrutiny. According to the motion, State Farm has referred to industry data, industry studies, and science in explaining aspects of its wind and hail practices. Yet plaintiffs contend that State Farm documents attribute at least one full-roof-replacement benchmark to Accenture Consulting Services and that it has not produced the underlying source data showing how that benchmark was derived.

Whenever an insurer says something is an “industry standard,” somebody should politely ask a few more questions. Whose standard? Based upon whose data? Covering which insurers, claims, locations, policy forms, and time periods? Was the data independently validated? Was the benchmark created to describe what actually happens in the marketplace, or was it created by a consultant as a target for what somebody believed should happen? Calling something an industry standard does not magically make it one.

Accenture, for its part, has objected to the subpoena on numerous grounds. It argues, among other things, that the information should first be sought from State Farm, that the requests are burdensome and overbroad, and that there are jurisdictional problems with compelling an Illinois non-party to produce the requested material. Accenture also says it did not work on the individual Hursh homeowners claim and performed no work on that action in Oklahoma.

Those arguments should be fairly considered. But the plaintiffs have a practical reason for wanting Accenture’s own files. A consultant may possess drafts, underlying data, internal analyses, workpapers, communications, and assumptions that never made it into the final presentation delivered to the client. Indeed, the plaintiffs specifically argue that some of Accenture’s source materials and internal work are not available from State Farm’s production.

Now let me turn to California, because one little breadcrumb in this motion should get the attention of lawyers handling State Farm wildfire cases. The exhibit index specifically identifies an “Accenture California State Model analysis.” The motion also describes the wind and hail work as sitting within a five-state program involving Florida, New York, California, Texas, and Michigan.

That does not prove State Farm used a similar indemnity-reduction program to handle California wildfire claims. I would not write that, argue that, or suggest that the Oklahoma evidence proves something it does not.

But I certainly would start asking questions. What was the California State Model? What did Accenture do in California? Were there other consulting engagements involving California property claims? Were financial “opportunities” identified? Were claim outcomes benchmarked? Were management approvals required for particular types of payments? Were adjusters measured for adoption of particular tactics? Did any similar methodology extend into wildfire claims?

The economics of wildfire claims are obviously different from hail claims. A wildfire program might not concern partial versus complete roof replacement at all. Pressure points could instead involve cleaning versus replacing smoke-damaged contents, remediation versus replacement of building components, the extent of smoke and soot contamination, additional living expense duration, code upgrades, debris removal, valuation issues, and decisions about whether the property has sustained covered physical damage.

I would be particularly interested in any California materials using terms such as State Model, Fire Model Enhancement, severity, leakage, opportunity, indemnity reduction, benchmark, claim quality, management approval, tactic adoption, or root-cause analysis. Discovery should not assume that somebody thoughtfully used the same program name in every state and for every peril. Corporate programs have a funny habit of changing names when they move around an organization.

The broader lesson is one I think property insurance lawyers need to appreciate as claims organizations become more sophisticated. Modern claims decisions are increasingly influenced by data analytics, consultants, centralized management, artificial intelligence, financial modeling, and performance metrics. The person whose name appears on the claim correspondence may be exercising judgment within a system that was designed years earlier by people the policyholder will never meet.

That does not make the system improper. It does mean we need to understand the system before deciding whether the individual claim received the fair and independent adjustment promised by the insurance policy.

For now, the Oklahoma filing gives us questions rather than final answers. It points to the next room, tells you which door may be worth opening, and occasionally leaves enough breadcrumbs on the floor that only a very incurious lawyer would walk the other way. The California State Model is one of those breadcrumbs.

Thought For The Day

“It is a capital mistake to theorize before one has data.”
—Sherlock Holmes, in Arthur Conan Doyle’s A Scandal in Bohemia

1 Hursh v. State Farm Fire & Cas. Co., No. CJ-2025-2626 (Okla. Dist. Ct. – Okla. County).