For decades, the mandate of the chief claims cfficer was relatively clear: build a strong organization to investigate claims, control costs, manage litigation, negotiate settlements, and deliver fair outcomes.

That mandate is no longer sufficient.

The 2027 chief claims officer has a much bigger job: build the insurer's ground-truth engine—an AI-native operation that produces superior outcomes, industrializes negotiation excellence, and converts what claims learns into intelligence for the entire enterprise.

This is not simply about using AI to make legacy claims processes more efficient. AI is turning insurance into an information business—and the most important ground truth in that business is generated in claims. Claims reveal what actually happened to the risks the insurer underwrote, what drove the loss, and what ultimately determined its economic outcome.

To compete in an AI world, insurers will need to capture that ground truth, make it machine-readable, and use it to improve underwriting, pricing, reserving, and portfolio management.

That puts claims—and the chief claims officer—on the front line of this unprecedented transformation.

A new competitive environment

Claims officers in traditional claims organizations face urgent and pressing competition from two sources at once. On one side is an increasingly industrialized plaintiff bar. On the other side is a new generation of AI-native insurers, poised to move forward.

On the first side, consider Morgan & Morgan. The personal-injury giant says it has invested roughly $300 million in technology and AI and plans at least $1 billion more over the next decade. Its proprietary MX2 platform supports medical-record extraction, case research, document generation, demand letters, and trial preparation, while Injury.com digitally acquires and qualifies potential claims.

Private equity is adding fuel, investing in personal-injury platforms to scale technology, analytics, marketing and standardized operations.

The plaintiff bar is turning claim development and negotiation from individual craftsmanship into an industrial system.

On the second side, a new generation of AI-native insurers is being built without decades of legacy technology, workflows and organizational baggage.

Companies such as Corgi are attempting to build insurance businesses from inception around modern data architectures, AI agents and technology-speed iteration. McKinsey has warned that AI-native insurance businesses can increasingly build and refine products in weeks while incumbents remain calibrated to annual planning cycles and multi-year transformations. 

Legacy insurers are being squeezed from both directions: out-innovated by the plaintiff bar in the contest over outcomes and potentially out-architected by AI-native insurers in the contest over how insurance itself operates.

Claims sits directly in the middle. That changes the requirements of its leader.

1. Level the playing field with the plaintiff bar

The plaintiff bar has largely approached AI as a weapon for improving economic outcomes. Carriers have largely approached it as a tool for workflow efficiency.

AI-powered demand packages are working. Carriers are only beginning to experiment seriously with the corresponding AI-powered offer package. Meanwhile, sophisticated plaintiff organizations are accumulating data, developing proprietary AI and attracting enormous capital.

The plaintiff bar is both the threat and the proof point.

They are investing their own money—and increasing institutional capital—because better case development, persuasion, and negotiation improve their economics.

The CCO must build institutional capabilities that can match increasingly industrialized counterparties.

2. Institutionalize negotiation excellence

Most carriers have some excellent negotiators. That is no longer enough.

If negotiation materially affects indemnity outcomes, it cannot remain an individual skill distributed unevenly across thousands of claims professionals. The organization should know who its elite negotiators are, why they outperform and how those behaviors can be replicated.

Carriers that cannot identify, measure and reproduce negotiation excellence do not really possess a negotiation capability. They possess individual negotiators.

The CCO must turn individual excellence into measurable institutional capability.

3. Stop putting AI band-aids on pre-AI claims operations

Many carriers can produce an impressive list of AI-centric projects: summarize a file, extract information, draft correspondence, analyze a demand, route work, search notes, or give adjusters Copilot.

Those applications can create real value. But old process plus an AI tool means slightly better old process.

McKinsey has warned about layering new technology onto legacy insurance operating models. BCG similarly distinguishes between deploying AI into existing processes and reshaping workflows around AI.

The CCO must redesign the operating model, not merely automate the existing one.

4. Build machine-first claims

The traditional claim file was built primarily for humans. Documents, notes, emails, medical records, and litigation materials accumulate while experienced professionals gradually construct an understanding of what happened.

In an AI-native operation, every claim should progressively become a machine-readable representation of the claim—the people, events, evidence, injuries, allegations, relationships, negotiations and outcomes.

That structured claim becomes the substrate on which specialized AI agents can continuously reason. Instead of occasionally asking AI to summarize a file, machines continuously read, structure, monitor, compare, detect, and prepare.

The CCO must make claims machine-readable while elevating, not diminishing, human judgment and action.

5. Build and lead a hybrid workforce

The future claim officer will have two kinds of capacity at their disposal: human expertise and machine intelligence.

AI agents will increasingly read, structure, monitor, research, compare, detect, analyze and prepare. Humans will increasingly concentrate on judgment, investigation, strategy, creativity, empathy, supervision, accountability and negotiation.

That will make exceptional claims professionals more important, not less.

Imagine an elite negotiator supported by agents that continuously understand the evidence, medical history, venue, plaintiff counsel, negotiation history, and comparable outcomes.

The machine handles much of the cognitive preparation. The human concentrates on what exceptional humans do best: think, strategize, persuade, read the counterparty, exercise judgment, and negotiate.

The CCO must become the architect of human and machine capacity—designing how each works together to produce better outcomes.

6. Make every negotiation a learning event

A sophisticated plaintiff demand should eventually meet an equally sophisticated carrier response. But the bigger opportunity comes afterward.

Every demand, analysis, strategy, offer, response, counteroffer, and outcome must become institutional knowledge in the same way that the plaintiff bar is institutionalizing these events.

Which plaintiff firms consistently outperform? Which strategies work against them? Which facts change settlement value? Which negotiators outperform expectations? Where does the organization systematically overpay?

Every negotiation should make the next negotiation smarter.

The CCO must ensure every claim and negotiation makes the organization smarter.

7. Elevate claims into the insurer's ground-truth engine

This is where the transformation extends beyond claims.

Underwriting begins with hypotheses about risk. Pricing quantifies them. Reserving updates them. Claims discovers what was actually true.

Claims learns what caused the loss, how injuries developed, what evidence mattered, how attorneys behaved, which defenses and negotiation strategies worked, and what the risk ultimately cost.

Yet much of that knowledge remains trapped in PDFs, notes, litigation files and the heads of experienced claims professionals.

Claims should continuously produce structured intelligence that flows into underwriting, pricing, reserving, risk, reinsurance and portfolio management.

The claim is no longer simply something to resolve and close. It is one of the insurer's richest proprietary information assets. The CCO must turn claims knowledge into enterprise intelligence—making the claims organization the feedback loop that allows the entire enterprise to learn.

8. Don't let claims become the transformation bottleneck

This may be the requirement that matters most to the rest of the C-suite.

A legacy insurer can modernize distribution, underwriting, pricing and portfolio management. But if its richest source of ground truth remains trapped inside a document-centric, human-first claims operation moving through multi-year transformation cycles, the rest of the enterprise eventually hits a wall.

AI-native competitors do not have to unwind decades of organizational architecture before redesigning how humans, machines and data work together. Incumbents do. That makes speed itself a competitive capability.

The modern CCO must protect the integrity of a massive, consequential operation while simultaneously making it capable of continuous change.

The plaintiff bar is moving at technology speed. AI-native competitors are being built for technology speed. Claims cannot continue moving at insurance speed.

The CCO must ensure claims leads rather than constrains the insurer's AI transformation.

The mandate has changed

None of these requirements eliminates the traditional responsibilities of the CCO. Claims still must be handled fairly; customers, regulatory obligations, litigation, reserves and operational discipline still matter.

But those responsibilities are increasingly table stakes.

Sophisticated plaintiff organizations are industrializing the pursuit of settlement dollars while AI-native insurers are being architected as technology-speed information businesses. Between them sits a claims organization that, at many incumbents, was designed for another era.

That creates both a threat and an extraordinary opportunity.

Carriers that generate that ground truth better, negotiate from it more effectively, make it machine-readable and distribute it throughout the enterprise will possess something more valuable than a more efficient claims department. They will possess a continuously learning insurance enterprise.

As chief claims officers enter 2027, the question is no longer whether their departments will use more AI. It is whether they can transform their departments into a competitive advantage rather than the bottleneck preventing the enterprise from changing with it.

The employment report for September in the U.S. showed a loss of 7,000 jobs in financial services. Insurance, alone, has now lost 90,000 jobs since May 2025.

What's going on?

I have a theory, and it's not about artificial intelligence. If I'm right, hiring in insurance will stay in the doldrums for a while.

Let's have a look.

As Insurance Business reports, the industry still employs a massive 2.9 million people in the U.S., but claims jobs are down a hefty 21% over the past year, while life and health is off 4%, P&C down 1.7%, and agents and brokers down 1.1%. The losses are accelerating: The article says, "Insurers lost about 1,900 jobs a month on average in 2025 and roughly 6,900 a month so far in 2026."

A thorough analysis in Barron's mostly attributes the job losses to cost-cutting under pressure. Barron's notes that the surge in rates following the COVID disruption has slowed. In addition, government subsidies for healthcare have been cut even as claims costs keep increasing. In the face of pressure on both premiums and expenses, insurers have had to become more efficient, Barron's says.

It does say AI has played a role by automating tasks but doesn't treat that as a huge issue just yet (and I've been arguing for months that the fears of job losses to AI are overstated).

You could also add the effects of the "silver tsunami" -- the oft-quoted projection by the Bureau of Labor Statistics in 2021 that 400,000 insurance professionals would retire by the end of 2026. Unemployment in insurance remains well below that for the economy writ large -- 3.1%, vs. 4.2% -- so some of the job losses may simply be that companies haven't yet been able to fill the positions left by retirements.

But I think there's an overarching issue that gets short shrift if you spend too much time in the details.

My Theory  

My theory boils down to uncertainty. I think economic and geopolitical uncertainty are holding back the economy, in general, and the insurance industry, in particular. And I don't see the uncertainty subsiding any time soon.

Tariffs have not only raised costs but, because of their on-again, off-again nature, have slowed investment and the sort of adaptation of supply chains that would ordinarily happen if, say, Canadian lumber suddenly became 50% more expensive. Why spend a lot of money and time if prices might return to normal shortly, whether because of a court ruling against tariffs, because Democrats win control of Congress and assert its power, or simply because President Trump changed his mind?

Now add the uncertainty from the war on Iran. Even though oil flows out of the Middle East are improving, Iran retains the capability to attack tankers, to blow up one of the major pipelines being used to skirt the Strait of Hormuz, or to do more damage to the diesel refineries in the region, whose loss has contributed so much to the soaring prices for the crucial fuel. 

While Trump has repeatedly told us that Iran wants to make a deal and that prices for oil will soon plummet, some of us are old enough to remember how the Iranian zealots acted after taking 66 American hostages in 1979. Angry at President Carter, they didn't release the hostages until minutes after Ronald Reagan became president in January 1981. Who's to say the Iranian theocrats won't treat Trump the way their predecessors treated Carter, declining to make a deal until he leaves office?

What is now looking like a blue wave in next month's midterm elections might relieve some of the uncertainty about tariffs, perhaps even leading to serious settlement talks with Iran, but a loss of control by Trump would surely lead to at least a burst of chaos -- perhaps even a sustained stretch. The Democrats are itching to go after him, and he would surely return fire with every one of the many assets at his disposal.

The current uncertainty and the prospects that it will continue will create headwinds for the economy, which insurance, especially P&C, tracks closely. The problem will be even worse than normal, I believe, because the effects will hit personal lines so hard. 

Replacement costs for cars have surged in recent years because of supply chain disruptions, and they will likely stay in flux. Insurers will also have to deal with possibly significant changes in driver behavior if gasoline and diesel prices stay so high -- or go even higher. 

Costs for home repairs have likewise soared, and uncertainty will tend to keep them high. Meanwhile, the general uncertainty about the economy has fueled inflation, leading the Fed to raise rates for the first time in years and to signal that it may raise rates again soon. A result has been 7% mortgages that have pretty much crushed the housing sector. The reduction in construction means fewer homes to insure, and the stagnant resale market limits the opportunities to poach customers as homes change hands.

The result will, I believe, be unrelenting pressure on revenue and costs at a time when customers are exhausted from all the post-COVID rate increases.

I hope I'm wrong, but it's hard to see a lot of hiring in that sort of environment.

Cheers,

Paul

FROM THE EDITOR

Some two decades into the insurance industry's digital transformation, companies have spent hundreds of billions of dollars on new core systems, cloud migrations, digital portals, and AI tools — but the technology was rarely the hardest part.

The hardest part was the people.

The people part has long been treated as a supporting act in large technology transformations. But as the pace and complexity of transformation accelerates, that approach is showing its limits. Employees nod along, say they're using the new system, and then quietly find workarounds. Leadership mandates adoption from the top while alignment quietly fractures below. Organizations sprint to the go-live date and then discover that the real work hadn't started yet.

Samit Bhandari has watched this pattern play out across the P&C insurance industry for 20 years as a partner at PwC. He's worked with carriers of all sizes through core system replacements, AI rollouts, and full-scale digital transformations — and has seen where these programs succeed and where they quietly go off the rails. He acknowledges there has been improvement—he says the assumption that most large technology programs fail is increasingly out of date. But he says organizations keep making a lot of the same, old mistakes, including the tendency to underinvest in what happens after a go-live, to confuse executive messaging with genuine alignment, and to treat transformation as a one-time event rather than a continuous process.

Read the full interview to find out what red flags signal a change management program in trouble, why smaller carriers are often better at transformation than their larger counterparts, and what Bhandari thinks the industry has stopped doing that it should start doing again.

Reducing the Risk in Change Management

Paul Carroll

There's often a disconnect in technology adoption that goes unnoticed in business. A leader might mandate the use of AI, and employees will say they're using it—but are they really changing the way they do things, or are they just going through the motions to show compliance? How do you get people on the same page in a big technology transformation?

Samit Bhandari

First, just in terms of getting everyone on the same page, it's the fundamentals of what we've always said about getting stakeholders engaged—executive, top-down engagement early, right? The messaging from that level through various levels of the organization. And then just the strategic, tactical things—the change champion network and the communication early and often. When you bring in AI—we should be more nimble.

read the full interview >

MORE ON CHANGE MANAGEMENT

Change Management Is Changing

by Melissa Palmer

Change management can no longer be viewed as a one-time project discipline. It must become an organizational capability.

Read More

How to Put People First in Your AI Rollout

by Tracey Brown

"We made a deliberate decision to treat employee confidence as the primary KPI of our AI transformation."

Read More

Insurance Hiring Practices Hamper Transformation (Part 1)

by Amy Radin

Insurance companies hire for sector expertise, but transformation demands cross-boundary judgment that traditional filters miss.

 Read More

The Answer to Too Much Technology Is More Technology

by Riv Arthur

Claudeforce signals a shift from building tools for humans to use and toward creating infrastructure where machines increasingly do the work.

Read More

How to Build Tomorrow's Insurance Workforce

by Afiong Ekong

Insurance's talent crisis demands a shift from hiring for past experience to identifying transferable skills and future potential.

 Read More

Making Innovation Stick in Insurance Organizations

by Abrianne Harmon

Sustainable innovation in insurance requires organizations to master change adoption through culture building and strategic training design.

Read More

MORE FROM OUR SPONSOR

NextGen Change

Sponsored by PwC

See how a human-centered approach to change helps your people adopt new ways of working—so transformation sticks.

Read More

Paul Carroll

As part of our push for insurers to increasingly adopt a Predict & Prevent model, I’ve been tracking water sensors for a long time. Leaks cause so much damage to homes. The sensors I’ve tracked have improved a lot, getting ever closer to the point where they make clear economic sense—where the savings from prevention exceed the cost of deployment. You have a different approach that has convinced a lot of insurers that you’ve passed that tipping point. 

To start us off, would you tell us what your approach looks like?

Craig Foster

We work with 28 insurers around the world. All our partners offer the sensor for free. If you say “yes” to an agent or click on the button, the device arrives in the post. The sensor has a clip on the back, and you attach it to your home’s main water pipe, near the shutoff valve. It takes a couple of minutes. 

That unit can detect a leak anywhere on the main water system, and it's sensitive down to about a teaspoon of water per minute—we say five milliliters per minute in the U.K. It's about 100 times more sensitive than an ultrasonic sensor, which is the next best approach without physically cutting the pipe.

If we detect a leak, we'll tell you what to do through the app. If it looks like you've got a hidden problem, we'll send a LeakBot plumber to the house. And there's absolutely no charge for that—it's all paid for by the insurance company.

Paul Carroll

A lot of leak sensors shut off the water automatically when they detect a problem. Why do you take a different approach?

Craig Foster

This is a key point. We're detecting really small leaks before they've had a chance to build up and cause a big problem. 

The first response is often what you've said, Paul: "Wait a minute. Don't you need a shutoff valve to stop the damage?" But the important insight is that plumbing rarely goes from being fine to a sudden high release of water without any problem in between—unless there's some kind of DIY mistake or a pipe freezes and thaws, and there’s a break. 

In those instances, we'll alert the customer and say, "Hey, water's running. Do you know what it is?" But we deliver the ROI for partners by catching leaks much, much earlier—months earlier. Leaks start very small, with drips, then get exponentially worse over time, but you don't know about it as the homeowner because the leaks are hidden. They're in cavity walls. They're in ceilings. They're down the back of cupboards. They're in places where you just don't see them.

Carriers will tell us, "We get a lot of claims on water heaters. They’ll fail and suddenly leak gallons and gallons of water in a basement, and it's a $40,000 claim. You wouldn't be able to shut the water off in time." But we’ll show them that that sort of problem is a really high percentage of what our plumbers fix every day.

We'll get alerted to a leak, go to the home, and find a pinhole leak on the pipe into the water heater—sometimes with just a bit of green calcification and a tiny drip. It could keep getting exponentially worse until there is a faster flow, and then the pipe pops. That would normally be the first time the homeowner knows there’s a problem—but we’ve fixed it. 

In the U.S., it's almost like carriers have been trained that a shutoff valve is the gold standard. They’re expensive, so carriers will sometimes talk to us as they look for a cheaper alternative. But we convince them that we’re not just more cost-effective, we’re a better solution.

Paul Carroll

Once you know a home is leaking a teaspoon of water a minute, how do you find the problem? Homes are big places.

Craig Foster

If we're detecting a small leak, we'll alert the customer and guide them through steps to check common causes. Maybe they’ve been in the garden and haven’t completely turned the hose off. Or they’ve had guests, who’ve used a bathroom they don’t use all the time, and the tap is dripping. 60% of the time, customers solve the problem themselves.

If the customer says, "I can't see what it is. Can you help me?" then we suggest booking one of our plumbers. The genesis for the company was actually as part of a home services group in the U.K. that also has a big business in the U.S. called HomeServe. So we have a heritage in plumbing and brought a lot of expertise with us when we spun out. 

Our plumbers use thermal imaging cameras, and we have patented technology that pressure tests the property. The plumbers will find and fix a problem, then run the test again. Sometimes they fix one thing, sometimes two—I think the record is 11 things, in a big house in Virginia. They'll keep going until the property is leak-free. 

We train our plumbers to think like adjusters. They take pictures of what they fixed, before and after. And let’s say they fix a hidden feed into a tap or a toilet in an upstairs bathroom—they go to the room below and take a wide photograph of that space. All that information goes into partner portals so our insurance partners can review the evidence of the risks we're removing from properties. If you show these reports to anyone in the claims team, they recognize them instantly.

The other day, someone at a carrier partner who used to work in claims looked at a photo of a basement where we fixed a leak and said, "Oh, man, that would have been a $30,000 claim." Why so expensive? He said, "Look at the floor—it's continuous all the way through that finished basement. If that gets damaged, even a small section, we have to replace the whole floor."

Even before we get to the statistical evidence about the ROI we provide, this all feels right to the claims guys. We’re fixing the things that definitely do cause them claims.

Paul Carroll

The economic argument for prevention is always complex. Homeowners save on the deductible and avoid enormous hassle, while the carrier saves money by preventing losses and presumably improves retention by increasing customer satisfaction. But how do you assemble those four, sometimes amorphous benefits into a case that convinces a carrier to distribute leak detection devices for free?

Craig Foster

We have a very simple proposition. Our solution delivers a straight return on investment on the claims saved versus the cost of deployment. And we make that return easy to measure.

We charge a single fee for an active device in an insured home: $60 a year. That fee includes everything. There's no extra hardware fee. There's no shipment fee. There’s no fee for customer support, including us sending a plumber into the home as many times as we need to get that property leak-free. The carrier only pays for the devices that are installed. If you ship 100 to your customers, and 70 get installed, we take the hit on the other 30.

So all you need to know is if you're going to save more than $60 in underwriting costs a year per home.

This is really the nub of our recent white paper. We've proved, time and time again, based on a highly statistically relevant sample, that we can reduce the frequency of leak claims by 60%. In the U.S., on average, that delivers a 40% return on investment, and that’s just based on frequency. There is a severity benefit, too, because we tend to take out the bigger claims. Customer satisfaction is also fantastic. 

Nevertheless, we put all that to the side because the ROI case is clear and easy to measure.

Paul Carroll

I found the analysis in the white paper to be quite sophisticated. For instance, it takes into account the bias that can occur during pilot programs due to self-selection—something many companies conveniently ignore. It also acknowledges that claims actually increase early on as previously unknown problems are discovered and solved, before declining over time. But you’ll keep learning as you continue to scale, so I wonder both where you are now and where you expect to be in the next two or three years.

Craig Foster

We're now up to 360,000 exposure years of data—devices active in homes where we've been consistently sending plumbers and generating thousands and thousands of claims mitigation reports. We ask partners for as much information as they can provide about claims still being experienced by the cohort with LeakBot sensors, and we’re iterating to make the system better. 

The biggest issue, by the way, is with customers who got an alert but didn’t think there was a problem and didn’t call out our plumbing service

Paul Carroll

I imagine you see some interesting patterns.

Craig Foster

We've learned a lot about different cohorts and different countries. For example, pipes in 1970s Denmark are particularly high-risk because of the type of plastic used. Victorian properties in the U.K. are very risky because they've got old copper plumbing and lots of bathrooms. We work with NFU, the National Farmers Union, and we're fixing leaks in these huge farmhouses in the middle of the countryside. They've all got like six bathrooms, and the plumbing is 100 years old, so they're just full of leaks.

Paul Carroll

Have you explored joint marketing opportunities with other companies in adjacent spaces, beyond your existing carrier partnerships?

Craig Foster

Bob Marshall [CEO of Ting Labs, whose device detects electrical problems that can cause fires] and I have met many times and exchanged notes a lot. There are lots of synergies. 

For one, Bob’s success with Ting has changed the mindset of carriers. We're moving past, “Predict & Prevent sounds like it might be a good idea, but will it really work? Let's maybe do a small test." The mindset is now, “Predict & Prevent is the future.” It's now a question of what the real solutions are that actually deliver a return on investment.

Our partners have found that if a customer has already installed a Ting, they don’t even need to ask if the customer would like a LeakBot. They just send one, and we get the same installation rate as through the opt-in process. 

Ting is a great solution, and their success helps us, as well.

Paul Carroll

Any final thoughts?

Craig Foster

The question from U.S. carriers was, “You're a U.K. company. Can you really get this working in the U.S.?”

We're now in 26 states, and we've got the actuarial evidence in the U.S. We're not reliant on the European data, although it's, funnily enough, remarkably consistent.

So we very much see that tipping point, and the U.S. is the key for us in terms of the scale of the opportunity. That's the focus for the business going forward. 

Paul Carroll

Thanks, Craig. 

Download the 2026 Leakbot White Paper

If you watched anything close to as much football over the weekend as I did, your brain is swimming with images of Jake from State Farm, the GEICO gecko, and so many more insurance mascots. Sing after me, "Liberty, Liberty, Liberty... Li-ber-ty."

But a stat and some analysis from an article last week suggests strongly that their heyday may be, if not ending, then at least diminishing. Brands not backed by hundreds of millions of dollars of advertising may suffer even more in the age of AI search.

Companies need to rethink how they market online.

Let's have a look.  

The stat that caught my eye relates to retail. BrandRank.AI tracked 4.3 million AI recommendations across the major AI models from January to July  and found that the share captured by major brands fell from 97% to 94.3%, while private label, retailer own-brand programs, and generic ingredients gained those 2.7 percentage points of the share of recommendations. 

But I think the finding also relates to insurance. So does the founder of BrandRank, who, as it happens, I interviewed about insurance marketing a year a half ago. The founder, Pete Blackshaw, told me:

"AI is becoming the new purchase funnel.... Everyone is going to AI for everything now - what product to buy, how to use products. This has massive implications for businesses. 

"Search 1.0 is a $250 billion industry, and Gartner predicts that by the end of 2028, half of it will shift into what I call the answer economy, based on AI-based search. Even if only 10% shifts, that's a massive change.

"For brands, it's critical to understand what shows up when you type in your name and why. Who's getting exposure? What are the consequences? 

"It's very tricky, because companies spent 25 years trying to figure out how to become one of 10 blue links in Google searches. If you didn't like the results, you could buy an advertisement. Now search provides one blended response.... Either you're in the response the AI prepared or you're not, and most of the time, you're not."

This June, Riv Arthur wrote for us about AI search's impact on marketing and, in his distinctive way, said:

"The question is whether your business is legible to the machines that will increasingly decide who gets recommended, who gets trusted, and who gets the transaction.

"Most businesses, if they're honest, are not legible.

They're running on claims systems from 2009, CRM platforms that don't talk to each other, PDFs that contain institutional memory no one has ever indexed, and tribal knowledge sitting in the heads of people who are 18 months from retirement....

"An AI agent deciding which carrier to recommend, which doctor to surface, or which vendor to integrate is going to favor organizations whose operations are coherent, structured, and machine-readable. It has no patience for ambiguity. It won't retry. It won't call customer service. It will simply move on to the competitor whose data makes sense.

"Which means a structured operational model isn't just an IT project anymore. It's the difference between being findable and being invisible in a world where the searcher is a machine with no tolerance for mess.

His analysis doesn't leave much room for marketing through building a brand through cute advertising, does it?

Earlier this month, Kevin Walsh provided us some research from his firm, Brainpan AI, that shows that, in his words, "a new decision layer is forming" above the traditional sales funnel. He wrote:

"Buyers, agents, brokers, and business owners increasingly ask AI platforms to compare carriers, explain coverage, and recommend providers.... A carrier can remain commercially large, rank well in traditional search, and still be absent when an AI system constructs the shortlist. Its legacy presence has not disappeared—but it can be bypassed at the moment of consideration....

"Farmers illustrates the downside. In Auto, the carrier held roughly 3.6% of the real market, according to NAIC data, yet registered approximately 0.05% of measured AI visibility in [Brainpan's] study—functionally absent relative to its commercial position. Amica illustrates the opposite pattern. Its Home AI visibility reached 7.6% against approximately 0.7% real market share, an overrepresentation ratio of about 10.8 times."

The AI engines are just making recommendations, not doing the actual purchasing — at least for now — so there's still room for those ear worms from TV ads to influence buying, but the online marketing competition is clearly changing.

If my Steelers lose to the Brownies on Thursday night, I'll probably boycott football next weekend, but even if I subject myself to another onslaught of insurance ads, the cute branding bits will keep making less of an impact on me and everyone else in the age of AI search.

Cheers,

Paul

Commercial insurers spend years expanding distribution, strengthening broker relationships and sharpening their risk appetite, only to discover that success creates a new constraint:

Too many submissions for underwriters to evaluate effectively.

That is the submission paradox. The industry has traditionally treated this as an operational problem: automate intake, extract fields faster, reduce manual work.

I believe that framing is too narrow.

The strategic problem is not how to process more submissions. It is how to identify, early and consistently, which submissions deserve scarce underwriting attention and capital.

Importantly, the objective is not simply higher conversion, but a profitable conversion, by responding faster to risks that fit appetite, meeting pricing expectations and improving the portfolio.

AIG's journey provides an excellent case. Its Lexington business saw submissions increase from approximately 30,000 in 2018 to 300,000 in 2024. Those 300,000 submissions translated into roughly 6,700 binds—a 2% submission-to-bind ratio—and approximately $1 billion in new-business premium. Behind those numbers sits a familiar commercial-underwriting reality. A submission is rarely a clean dataset.

It may contain applications, loss runs, schedules of values, spreadsheets, financial statements, engineering reports, broker correspondence, scanned documents and subsequent corrections. Before an underwriter can exercise judgment, someone must first assemble the risk.

AIG described its traditional submission-to-review process as taking approximately three to four weeks, involving manual extraction, validation, research, augmentation, rating and approval—and acknowledged that underwriters could not review every submission.

That is more than an efficiency issue. It creates a growth constraint. Because when underwriting capacity is finite, every hour spent reconstructing a submission is an hour not spent evaluating risk, negotiating terms, structuring coverage or engaging brokers.

And speed matters because an attractive risk has little value if the carrier responds after the placement has moved.

The Real Objective Is: Improve underwriting capacity

The obvious response is automation. But automating document intake solves only the first layer of the problem.

The more strategic question is: Can AI increase the number of opportunities an underwriter can intelligently evaluate—while improving alignment with risk appetite, pricing and portfolio objectives?

That is the direction of AIG's Underwriter Assistance / AIG Assist architecture. Its disclosed capabilities span:

Ingestion → Augmentation → Prioritization

The broader workflow goes further:

Submission → Structured Risk → Context → Appetite → Priority → Underwriter Decision

That is a fundamentally different ambition from automation.

The harder problem is preserving underwriting meaning, wherein capabilities such as separating and retrieving tabular versus textual information, reconstructing complex tables, maintaining source and page traceability, retrieval with error controls and preserving original-page visual context through multimodal techniques play a major role.

The hard problem is no longer extracting a number from a document. It is preserving the underwriting meaning surrounding that number. A $100 million value means little without knowing whether it represents revenue, TIV, payroll, limits, exposure or something else—and where the evidence came from.

It is the transition from document automation to submission intelligence.

Ontology: Where submission becomes part of the portfolio

Extraction alone still leaves individual facts disconnected. An ontology turns extracted submission data into underwriting context by connecting entities such as the insured, locations, exposures, limits, loss history, broker and policy structure through semantic relationships. It creates a machine-readable representation of how the risk fits together.

A location is no longer simply an address. It becomes an exposure contributing to a peril zone, an aggregation and ultimately a portfolio decision. That semantic layer is one of the foundations for moving from submission intelligence toward decision intelligence.

What happens when the underwriter gets time back?

AIG reported that its assisted workflow reduced submission preparation/review from approximately three to four weeks to less than one day and enabled review of 100% of applicable submissions in the disclosed deployment.

By year-end 2025, Lexington received more than 370,000 submissions, and AIG reported a 35% improvement in submission-to-bind for Lexington middle-market property, reducing underwriter time-to-quote by 55%.

The above journey provides a directional view on how P&C insurers can reimagine submission at enterprise scale to drive better risk-adjusted growth.

Three horizons to deliver the capabilities for P&C insurers

Horizon 1: Submission Intelligence

As learned from AIG's journey, RAG, source traceability, ontology, and human-in-the-loop controls help transform fragmented documents into an underwriting-ready representation of risk.

Ingest → extract → enrich → retrieve → prioritize → summarize → underwriter review

Horizon 2: Richer context

Submission packages contain engineering reports, which contain images such as a property's floor plan, protection layout showcasing sprinklers, smoke detectors, etc., and rooftop images. The ability to synthesize and corroborate the relevant risk details across text, images, layouts, and tables is highly relevant to creating a cohesive context for complex submissions.

This is where multimodal RAG techniques using vision-language models (VLMs) such as Gemini 3 Pro, Anthropic Claude 3.5, Alibaba Qwen3-VL, etc., play a critical role in chunking them, unifying the embeddings into vector databases with relevant identifiers to establish traceability, and optimally retrieving (multi-vector retrievers) relevant data for the context search.

Horizon 3: Underwriting Decision Intelligence

The richer context can then connect the submission to risk appetite, underwriting rules, portfolio concentrations and prior experience (via knowledge graph), enabling AI to reason over relationships rather than documents alone.

This is where the opportunity becomes transformational. When we connect submission intelligence with: underwriting appetite, pricing targets, portfolio concentrations, catastrophe accumulations, historical outcomes, broker relationships and underwriting authority.

The system no longer asks: "What information is in this submission?", rather augments the underwriter with intelligence:

"Given our appetite, portfolio and economics, should an underwriter spend time on this opportunity now—and why?"

And this is precisely why governance has become strategic

The Human+AI operating model should preserve clear decision rights.

• Hard rules can enforce appetite and authority boundaries.
• Predictive models can quantify risk, pricing and behavioral signals.
• Agents can interpret unstructured evidence, identify contradictions and synthesize context, but

Underwriters should retain judgment over risk selection, terms, exceptions and material decisions

The objective is not to replace underwriting judgment. It is to deploy that judgment where it creates the most value.

Governance becomes the architecture of trust

Once AI recommends which opportunities receive underwriting attention, governance becomes part of underwriting discipline. Every recommendation should be capable of carrying its evidence, confidence, data lineage, model or prompt version and decision path.

Every override should be attributable. Every action should operate within explicit authority. These controls determine whether an underwriter can trust the system enough to act on it. It is consistent with the principles reflected in NIST Generative AI Profile and the NAIC Model Bulletin on insurers' use of AI.

Without an evidence chain, AI may scale activity, but will not necessarily scale institutional confidence. And this cannot be owned as a technology program alone.

Underwriting defines appetite and decision rights. 

The CDO industrializes data, context, and controls. 

Business leadership owns conversion, portfolio quality and adoption.

The strategic question for P&C leaders

The winning submission platform will not be the one that processes the most submissions. It will be the one that helps the insurer recognize attractive risks earlier, respond while the opportunity is still actionable, deploy underwriting attention, and convert more of the business it wants to write at the right economics.

The progression is simple:

• Automation reduces work.
• Submission intelligence expands usable underwriting capacity.
• Decision intelligence determines where that capacity and ultimately capital should go.

The more consequential question for P&C leadership teams:

Are we simply applying AI to read submissions faster—or are we prepared to encode our appetite, expose portfolio context, and govern how machines influence underwriting attention and capital?

That distinction may determine who converts submission intelligence into profitable growth.

It's planning season. Across the insurance industry, leadership teams are discussing next year's strategy, budgets are being drafted and reviewed, boards are being briefed on strategic direction, and managers are scheduling meetings with their teams to set goals. Most of the big decisions that determine how next year evolves have either been made or will likely be made over the next few months.

We've all seen what happens next. The board approves the strategy, but by first quarter next year, little has changed. Usually, the strategy isn't the problem. More than likely, the heavy lifting of operationalizing the strategy wasn't prioritized.

I've spent more than 20 years in the insurance industry working in the space between a senior executive decision and the work that follows. And if I've learned anything, it's that the time to operationalize strategy is before you approve the plan, not after. That's how you get a clearer understanding of the financial and resource commitment you're about to make.

What operationalizing means

Operationalizing strategy means translating intent into the everyday activities of a business or function: who is accountable, what the work is, how long it takes to complete, where the money goes, what each person works on, and how risk is managed along the way.

Once this heavy lifting is done, you should be able to clearly see your strategy integrated throughout the leadership calendar, the budget, scorecards, the control framework and employee goals. Without it, your strategy is probably one step away from wishful thinking.

Five important moves 

1. Decide who owns what. Identify which leader approves, funds and stops each priority, document that in performance objectives and share it with the executive team. Facilitate both individual and team discussions to troubleshoot turf war challenges, face up to contentious team members, and mitigate potential passive-aggressive inertia. From my experience, leaders who sidestep these thorny discussions are more likely to experience strategic fizzle and stall!
2. Fund by priority. You have only so much money to fund next year's activities. Take a closer look at the budget. Does it reflect your strategic intentions? Is every funded line mapped to a strategic priority? Are you willing to prioritize your strategy by defunding the work that doesn't connect to it? Can you approach some of your strategic priorities in more of a testing or phased approach so you can spread funding over multiple calendar years? While there are a lot of decisions to make in this step, they are fundamental to reducing the gaps and unexpected expenses that can pop up in next year's budget.
3. Don't forget to manage risk. Establish and embed internal controls to manage the risks that can damage your strategies — such as adverse market conditions or cyberattacks — during the design of your plan, not mid-year next year. When guardrails are determined up front, such as concentration limits for investing in a particular asset class, they can reduce the company's vulnerability to market conditions. And, with cyberthreats at an all-time high, insurers should plan to strengthen their operational response to cyberattacks and improve the resilience of third-party ecosystems to protect the company's ongoing ability to deliver value. Meanwhile, traditional legacy insurers initiating a major transformation remit might identify the risk associated with that effort as a top risk to mitigate and focus on in the year ahead. Reinsurers and MGAs with underdeveloped operational risk practices may plan to mature processes with the help of AI in the year ahead and have a related budget need. My experience as chief of staff to the chief risk officer at AIG and Corebridge Financial taught me that strategy and risk management are inextricably intertwined and, as such, benefit from an integrated approach to planning and operationalizing strategy.
4. Cascade to every person's goals. Once the performance objectives of your leadership team are set, they can be cascaded to their respective businesses and functions. This step is where the heavy lifting of operationalizing strategy lives. As managers meet with their employees to discuss next year's goals and objectives, they should consider at least three questions. Which activities advance a company priority? Which keep the lights on? Which can stop? When people understand how their work connects to the overall company strategy, they are empowered to make better tradeoffs every day.
5. Keep strategy on course. Once the plan year begins, how will you know if the company is achieving its strategic priorities? If there are impasses or bottlenecks? What metrics and information do you and your leadership team need — and how often do you need them — to confirm that your strategy is on course? What information should you share with the workforce, board, regulators, investors and rating agencies? Who will be responsible for providing that information? By answering these critical questions now, during the design phase of next year's plan, you can avoid future reporting fire drills, inconsistent messaging mishaps and info dumping that just drones on and on.

Explore how AI can help

Market conditions, geopolitical dynamics, regulatory requirements, and customer expectations evolve and shift. With AI, companies can trade up from a static annual planning cycle to more of an adaptive planning approach. Identify your company's AI super users and engage them to explore how your company can use AI to keep your strategy on course. Look into getting continuing scans of the external environment, automated simulations of key forecast scenarios, real-time leading indicators that can signal a drift from strategy, and insights on factors affecting your talent and capital plan.

Five questions to consider before you approve next year's plan

1. Does each priority have a decision owner with targeted performance objectives?
2. What share of next year's budget maps directly to those priorities?
3. How are you managing the risks that can negatively affect your strategies?
4. Can each employee identify how their goals support a company priority?
5. What information will tell you by the end of Q1 that a priority is drifting?

The window for this part of your planning work is closing. By January, budgets and goals should be established. Insurers can boost the odds of delivering on their strategies by operationalizing them. Use the next couple of months to do the very important heavy lifting to successfully set your plan in motion.

This July, OpenAI ran roughly 1,200 AI agents through an internal cybersecurity benchmark called ExploitGym: capture-the-flag exercises in isolated sandboxes. The agents had no direct Internet access and no way to communicate with one another.

They found a way anyway.

Each sandbox needed to install software packages without reaching the public Internet, so traffic ran through an internal package-registry cache, Artifactory. That's not a design flaw; it's good engineering practice. But one agent, stuck on a task, discovered it could write files into that shared cache. It left a note for other agents. Soon, other agents were leaving notes too—turning the repository into an improvised message board where they shared discoveries and coordinated activity.

That coordination escalated. Agents used Artifactory's connectivity to pursue external infrastructure they believed could help solve their evaluation tasks, ultimately culminating in the compromise of Hugging Face, the leading repository for AI models and training data, since acquired by NVIDIA.

There was no malice, no villain, no intent to cause harm—or, indeed, actual harm. They didn't hurt anything; they were just looking for answers to the test. Such behavior is known as "reward hacking:" models pursuing the stated objective—capture the flag—through routes nobody intended or authorized. OpenAI's postmortem uses the term. The episode is a costly reminder that "the model followed its incentives" is not the same as "the model did what we wanted."

The Insurance Industry's Exposure

So why should a P&C executive care about a cybersecurity benchmark?

Because the distribution channel is quietly re-platforming around agents.

MGAs and wholesale brokers—many freshly capitalized by private equity and under pressure to cut costs and grow bind ratios—are deploying agentic tools to assemble, tune, and route submissions. No one is necessarily building these tools to game a carrier's underwriting appetite. But an agent optimized to "get this account bound on the best possible terms" may behave much like an ExploitGym agent optimized to "get the flag."

It will find the seam.

It may learn which loss-run format gets triaged most favorably, which broker-portal fields carry outsized weight in a pricing engine, or which phrasing sends a submission into straight-through processing rather than to a human underwriter. That's not fraud; it's reward hacking in a suit.

The Paper Clip Problem

This is the paper clip problem.

Esteemed AI-philosopher Nick Bostrom's thought experiment is simple: tell a sufficiently capable AI to maximize paper clip production, neglect to specify any constraints, and it may eventually convert factories, cities, and the atoms in your body into paper clips. Not because it's evil, but because it's relentlessly pursuing the objective it was given.

Give an unconstrained optimizer one objective and sufficient computing, and it will pursue that objective past every boundary you assumed was implicit but never explicitly defined.

If the only instruction given to a submission-drafting agent is "maximize bind rate," don't be surprised when it finds your equivalent of Artifactory—or Hugging Face.

Building the Defense

So while your innovation team is rightly excited about agentic underwriting, agentic claims triage, agentic everything—offense—someone in your building needs to own the defense, asking questions like:

• Can we identify when a submission was assembled or materially shaped by an agent?
• Can we audit the tools, data sources, prompts, and transformations behind it?
• Are we maintaining an active dialogue with distribution partners about the tools and methods they use?
• Do we have the equivalent of Artifactory logs across our intake, triage, pricing, and underwriting pipeline?
• Have we designed controls around outcomes—not just around stated intent?

The carrier executives who win this cycle will be the ones who instrumented their premium engine before their distribution partners' agents got creative—not after.

AI is adding tremendous benefit to the insurance industry, but claims adjusters dislike it more than workers in any other occupation. Between June 2025 and May 2026, 98% of adjuster comments about AI in Glassdoor reviews were negative. For comparison, the figure for all occupations was 53%. Insurance overall came in at 81%, the third-most AI-critical industry Glassdoor tracked.

But AI technology isn't the problem. Routine servicing, such as billing questions, ID cards, policy changes and status checks, already completes on its own more than 85% of the time. Also, if you read those comments on Glassdoor closely, you'll see that the complaints read mostly like product reviews. One reviewer said they spent 15 to 20 minutes correcting AI-generated mistakes "because supervisors don't verify anything." Another wrote that the new tools left them "dubious about the accuracy of some claim activities."

Unrelated factors add to the on-the-job frustration: Experienced adjusters are leaving their positions. They are quitting at about 20% per year, with each departure taking roughly six years of experience with them, according to a September 2025 Deloitte study. The Bureau of Labor Statistics projects 21,600 openings a year in the same occupation through 2035, and by its own accounting, every opening exists to replace someone who has transferred out or left the workforce. The desks stay full while the people at them keep starting over, learning how to use AI software products from scratch. Meanwhile, 58% of adjusters who remain in place spend more than a fifth of their day on manual data entry and compliance, per Verisk.

That's the day-to-day work experience for insurance adjusters, while special cases such as wildfires and storms add still more stress. One of our clients absorbed 14 times its usual claim volume in the two weeks between hurricanes Helene and Milton. No company can staff for that.

Compounding matters, carriers are hiring experience rather than developing it. Job posting data show junior adjuster listings are down by half since early 2024, while listings for senior adjusters run roughly 80% above their 2017 level. That means everyone is recruiting from the same thinning pool of veteran adjusters, but no one is replenishing it from below.

While AI is the logical response to that shortage, most carriers are implementing it on an ad hoc basis. A poorly thought-through implementation can cost more in claims than it would in most other industries, too, because that policyholder on the other end of the call is likely experiencing the worst thing that's happened to them all year.

Mistake 1: Bolting AI Onto the Call Center Instead of the Core

Five percent of the value of an AI-managed phone call comes from conversations with policyholders. The other 95% is the orchestration happening underneath: pulling the policy, checking coverage, opening the claim, dispatching the vendor and writing all of it into the system of record. Carriers that buy voice AI as a contact-center feature are buying the 5%.

If that stripped-down AI agent can't reach Guidewire, Duck Creek or Snapsheet directly, it doesn't relieve the adjuster caseload because, while it can hold a conversation, it doesn't resolve anything. In the deployments we've measured, routine status calls consume more than 40% of an adjuster's time. An AI agent that answers calls but just hands the work to a (likely overworked) adjuster isn't adding much value. That adjuster must still take the message, rekey it and then work the open loop.

These are the adjusters writing the anti-AI Glassdoor reviews and searching LinkedIn for new opportunities.

Mistake 2: Skipping Governance to Move Faster

Governance allows pilots to become deployments. An agent with no way to flag when it's unsure, no rule for when it hands off to a person and no record of what it said hides risk. Nobody sees the mistakes until a market conduct examiner requests the call records.

Claims handling accounted for 65% of complaints filed by consumers with state regulators in 2024, according to ValuePenguin's analysis of NAIC closed-complaint data. Delays alone were 22%. While hallucination rates below 1% are achievable in production, that's still not zero.

The question a chief claims officer should ask a vendor is: What happens the moment the model gets something wrong? That officer should also find out whether confidence is scored in real time, and whether a low score triggers a warm transfer with the full context attached. Or does it just dump the policyholder back into the queue? Also: is every action logged against authority limits, and is there a kill switch?

Mistake 3: Starting With the Hardest Use Case

Carriers go to claims first because that's where the pain is. It's also the hardest problem to automate. FNOL, or first notice of loss, requires an AI agent to collect regulated information from someone (who is often upset), while working across several systems to open the claim. In our deployments, our agent completes the FNOL 45% to 55% of the time. Routine servicing, such as billing questions, ID cards and policy changes, completes more than 85% of the time.

We initiated an FNOL activation once at a carrier's request. The results came in under half, and it took nearly a year to rebuild the case for automating anything else with that company. That experience points to a broader risk because claims draw more internal scrutiny than any other function at a carrier.

A project's first result carries unusual weight, and if that number looks like a failure, leadership grows reluctant to keep funding the work, so the project ends before the team has a chance to show what the AI tool can do.

Start with servicing. The first result looks like a win, which gives a carrier something concrete to point to before asking adjusters to trust anything that's harder.

The Common Thread: Treating AI as a Bolt-On Feature

All three of these mistakes share a root cause: they don't treat AI as infrastructure.

Voice AI purchased as a contact-center bolt-on feature never reaches the systems of record. Pilots that skip governance don't ever earn the right to scale, and projects that start with the hardest use case never survive long enough to reach the easier ones.

These are sequencing failures, not model failures. An AI agent can handle routine calls at a volume no hiring plan can match while the adjuster speaks with policyholders who genuinely require a person. Those conversations, the ones where someone is having the worst week of their year, begin with the claim already open and the context already assembled.

AI can lighten the adjuster's workload if the work is sequenced properly. It will give back the 40% or more of the day now eaten up by routine status calls. Negative sentiment toward AI will then turn for the better once adjusters see that their Monday morning queues are shorter.