FM Using AI to Elevate Claims to Deliver More Than Just Cost Savings

At FM, claims managers and claims professionals are taking a “trust-but-verify” attitude about the integration of artificial intelligence into their operations.

A data-driven, results-based approach makes sense for a company that touts its specialty as being an engineering-driven property risk management and loss prevention insurer for large, commercial companies. The a Fortune 500 company headquartered in Johnston, Rhode Island, has more than 6,000 employees worldwide.

So far, it’s hard to argue that the integration of AI into the company’s claims functions hasn’t produced numbers that point to greater efficiency, with a measured reduction of nearly 50% from report to date of payment on small losses.

To get a good handle what the company is doing with AI, Claims Journal spoke with Jeremy Gallant, senior vice president of claims at FM. Gallant has been in the insurance industry for roughly 25 years, nearly 20 of which have been in various leadership roles at FM.

The conversation has been edited for brevity and clarity.

Claims Journal: What new technology are you integrating into your claims department/processes?

Gallant: We’ve done quite a bit. First one we’ve done is called the FM Portal. This is a secure digital platform interaction between clients, brokers partners and claims professionals. And it’s really summarizing all of our claims information, our engineering information, and underwriting so you can report a loss. You can put your banking information in, then you can request payments and advance payments. You can upload documents. That’s been a big investment.

The second one, which is quite interesting as a not fully automation, but a pretty much automation of fast-track claims…that is currently losses under $150,000, but we are looking to increase that in the short term. And this (technology) is that once things get recorded into the portal, it will automatically follow up, it will automatically request the documents that are needed. A human is still in the loop on the back, reviewing those documents, but at a very quick and high level. We’ve seen some big improvements from that.

And then in terms of other AI tools we were using or planning to use, we are looking to analyze location data, loss data and documents related to loss. So, we are working on a project to better integrate claims data into engineering and reinsurance and underwriting and making sure all of those data sources between all those groups are going back and forth. We are looking to automate filing of documents and a lot of our statutory files. You might be surprised how much time is actually spent getting those into the right codes. So, we’re looking at how do we automate that, capturing data from different documents.

We started with OCR type technology and are kind of evolving from there, and we are currently in the pilot stage for using Gen. AI for reviewing our operating requirements, doing lease contract reviews and writing executive summaries and file narratives. We are also looking at kind of loss triaging, how to get better at that and certainly on the subrogation and forensic side, as well forensic accounting, looking to figure out how can we use some of this to eliminate either consultants we’re hiring or get better targeting where subrogation opportunities may exist.

Claims Journal: Why are you implementing this technology? What problem is it solving or will it solve?

Gallant: We’re looking at it from the view of what our customers asking for. So, on the small end of size, these are small, routine losses, we want to be efficient on those. But we take big, big shares of risk, and we have big losses as a company. So, we approach it as we want to get the claim as accurate as possible, doing the most limited amount of work possible.

So, we’re looking at how do we remove the admin type of steps within the claims process in order to be efficient on the small stuff. And because we partner as a mutual company with our clients on larger losses, we want to be more strategic with what our adjusters are doing with them. We’re freeing up time so they can spend more time with our clients pre-loss, more time with our underwriters making sure the policy is clear, doing policy workshops. And when we do have big losses, we have more time to spend to get onto site quickly and help them mitigate their loss.

Apart from going through the claims process, a big, big part of what we try to do is (find ways to) get our clients back into business as quickly as possible to their benefit, certainly to our benefit, as well, but how do we make sure their customers aren’t going to be impacted as much as possible. We’re looking to drive down cycle times to get people paid faster, get the least amount of information we need to adjust the file properly, and again, allowing all of that to happen within the portal technology we talked about.

Claims Journal: How is it working?

Gallant: It’s been great actually. It’s moving at a fast pace. Some of these projects we’ve thought about starting last year, the technology didn’t really exist to make it possible. But this year, the technology has caught up, and we are able to do what we had planned last year. We’re seeing results for everything we’ve done. We’ve globally scaled our fast-track product. These AI tools are starting to work and we’re starting to see some measurable impacts on where our adjusters are spending their time, how many files they’re able to handle, and really how do they be more value-add and strategic on the on the losses by removing those admin type tasks.

Claims Journal: How long has it been up and running?

Gallant: The portal’s been up since April 2023 and the automated technology for fast tracks has been running about 18 months. It started in the U.S. and Canada and most recently expanded to Australia, the U.K. and the rest of Europe. And a lot of the AI tools, we started some work with in 2025. Some of that stuff was quite successful, and then this year, we’ve got a lot more. And I’d say (they are in the) advanced pilot stage and nearing implementation towards second-half of the year and certainly into 2027.

Claims Journal: How do employees feel about it?

Gallant: A great question. I think there’s excitement, some eagerness on one side. And on the other side, there’s some reservations…so I think it’s a little bit of all of the above, and I think people are trying to determine how does this affect their roles going forward.

As with anything new, there’s a period of transition that it takes the time to trust the technology and we’re a trust-but-verify type of company, so there’s resistance a little bit, but openness to using it and see how effective it can make their jobs.

I’d say, generally younger generations, they’re kind of quicker on the curve with this—they expect this from companies. And more experienced employees, some of them are just taking a little bit longer, but we’ve done a lot to bring on training around all of this AI. And as they start to use it, they’re getting more comfortable with it and they’re seeing results and it’s allowing them to do things that they enjoy, which is not the admin side of the job, it’s more consulting with clients and working with them to resolve issues and get them back in business as quick as possible.

Claims Journal: How do you feel about it?

Gallant: It’s been great. I think it’s been a learning curve for everybody. It’s an opportunity to relook at a lot of these processes that existed for a long time. So, it’s giving us the opportunity to take a fresh look at how we do things, what loss trends we’re seeing in the industry and how best we can respond to our client needs in the future. And some of the time, it doesn’t take that long to get some of these up and running—tand the impact you see right away.

I think it’s giving claims a great spot: to be elevated in companies and to deliver more than just cost savings focused on client value, (to deliver) improved client experience and being a better source of data that could be used back within the underwriting side, reinsurance side and engineering side, so ultimately we can better manage risk with our clients, we can better differentiate pricing and help our clients be more resilient and hopefully loss-free in the future.

Claims Journal: How do you measure the success of this technology? What are the metrics?

Gallant: I think it comes back to that (being) efficient where you need to, being and effective where you need to be. On our employee side, we’ve done satisfaction surveys and have gotten results back. That’s been generally positive. We’re measuring times that people spend on various tasks and looking at, ‘OK, if we can automate some of these tasks, how much time are we saving that can be used elsewhere?’ We’re seeing forensic savings and external loss adjustment expenses, how much we know, how much we’re getting back on subro recovery. We can track how we’ve improved in that regard and ultimately the number of files people are able to handle. So, we have a good sense on the employee side.

And on the client side, we measure our cycle time. In small losses, we’ve seen a reduction of almost 50% from report to date of payment. We’re working now on some surveys on the client-side experience. We certainly have a lot of verbal feedback that’s been greatly positive, but we’re going to firm that up into a little bit more data-driven (information). And then we’re working on measuring how our adjusters are able to help mitigate losses and what does that actually mean in terms of dollars saved or ability of our clients to get back into business sooner. So, I think there there’re lots of benefits and lots of things that are happening and we’re trying to capture all of them that we can.