At ITC Vegas 2026, Marina Barg, Chief Claims Officer, U.S., at Ascot Group, joined Jermaine Everett, Head of Sales at Bevaya, for the session “Taking Action in Claims with Agentic AI.” Ascot is a global specialty (re)insurer, and the session offered a practical look at AI in insurance claims, covering how its claims team got started, what got in the way, and what comes next.
Ascot learned that automating first notice of loss (FNOL) in specialty claims takes a partner that knows specialty, a clean process, and a claims-led rollout. Bevaya’s FNOL AI agent now runs at an 85% no-touch rate for Ascot and has cut the time to open a claim from 2 days to 20 minutes.
The full session is below, followed by the main points in Marina’s own words.
Read the full transcript and jump to any chapter: AI in Insurance Claims: How Ascot Started with FNOL Automation
Why Ascot Needed a Partner That Knows Specialty Claims
Ascot writes specialty lines such as ocean and inland marine, environmental, space, and kidnap and ransom, and each line of business has its own handling rules. Marina described Ascot as “more of a severity company.”
That shaped what Ascot looked for in a partner. “We want to make sure we partner with people who understand specialty claims and the nuances of it,” Marina said. “It was really important for us to find a partner who grew up in insurance.”
Ascot also runs an AI center of excellence, so every project starts with the same decision. “It’s always a question that we ask ourselves, do we build? Do we buy?” Marina said. Building isn’t a one-time effort. “Whenever you build something yourself, you have to not just build it, but you have to support it and innovate in that regard and keep on top of the most current technology,” she said. “Sometimes that’s the way to go, and sometimes we’re better off with partnering with experts in the field who know what they’re doing.”
For FNOL, Ascot chose a partner with deep insurance knowledge that could advise on the approach, not just implement it.
Why FNOL Automation Was the Right First Step
Ascot’s technology transformation started in claims with Guidewire, and FNOL was the first AI workflow on top of it. “We started with FNOL because that’s kind of the part that made the most sense, and quite honestly, it’s one of the ways that we can test you as a partner,” Marina said. “If all things fail, we could always go back to our existing process, and luckily it didn’t fail. It succeeded very much.”
The need was clear, too. Ascot wanted faster intake without adding people.
It also lays the groundwork for what comes next. An AI agent that can tell an FNOL from everything else in the inbox is the foundation for indexing the rest.
Starting with FNOL also gave Ascot a chance to look hard at the process itself.
“It really doesn’t make sense to automate a process that doesn’t work or a process that requires a lot of workarounds.”
Marina Barg, Chief Claims Officer, U.S., Ascot Group
How the FNOL AI Agent Handles a Crowded Claims Inbox
“We have a very lean group that manages our FNOL through an email box,” Marina said. An FNOL there “could be a three-line note from a broker saying, ‘This is the policy number. We have a claim with a one-line description. Please set it up.’” The same inbox also receives claim updates, documents, and status questions.
Bevaya’s FNOL AI Agent works that inbox in three steps: it spots the FNOL, applies the rules for that line of business, and creates the claim in Guidewire. Anything that falls outside the rules goes to a person through human-in-the-loop review. Marina described the effect for Ascot’s insureds: “They could get their claims set up a whole lot faster, acknowledged a lot faster, triaged to the right adjuster.”
If accuracy is the main concern for your team, our post on whether you can automate first notice of loss without sacrificing accuracy covers how that works.
What Made the Implementation Hard
An audience member asked about the challenges. “Some of it was our internal plumbing,” Marina said, from pulling the right policy for the right date to getting Guidewire the elements it needs to set up a claim.
The other piece was defining what counts as an FNOL in a mailbox full of other mail, and keeping a record of what happened to everything else. That record matters to regulators. “If we can’t explain it, if we can’t provide trace and audit record of what happens there, we’re going to have a problem,” Marina said.
Many states have adopted the NAIC model bulletin on insurers’ use of AI, which calls for a written AI governance program, and every Bevaya AI agent keeps an audit log of each action it takes.
Expectations took work too. “We expected to turn on the button once everything was done and things load perfectly, and they didn’t,” Marina said. Ascot launched in February and tracked the success rate as it went. “We knew that it was going to be at least a six months process for us to train the model in a way where we have confidence in it.”
How Ascot Brought Its Claims Team Along
Change management was a lesson Ascot first learned with Guidewire. “You get people who are so set in their ways for whom changing minor aspects of their workflow just creates, you know, ‘No, I’ve never done it this way,’” Marina said. For the AI work, Ascot kept claims in charge: “This was really a claims and business-centered and business-led initiative. It wasn’t IT telling us what to do.”
Ascot’s own claims people worked through the process with Bevaya to figure out what worked best, and results did the rest. “Once people see the results and can look past the initial ramp-up and the training of the model, and see the results and see the output, it was easy to get them to come along,” Marina said.
Marina was clear about the purpose. “We’re looking for solutions that will get our people to be able to make the right decisions faster. That’s ultimately, to me, the purpose of AI.”
What Comes Next for AI in Insurance Claims at Ascot
Ascot’s next step is claims indexing for the supporting documents that arrive after Ascot sets up a claim. “It’s not enough to just set up the claim and route it to adjuster. We can probably do better,” Marina said. The aim is to “be able to tie any of that to an existing claim, kick out anything that requires human intervention,” and identify what each document means for the claim.
On whether AI will soon handle claims end to end, Marina answered false. “First of all, the regulators will never allow it,” she said. “There’s always a piece of judgment in the decisions that we make.” On whether insurance regulation will catch up with AI, she also answered false: “Technology, I think, always outpaces regulation.”
Ascot’s experience shows that the hardest parts of claims AI sit outside the AI itself, in the plumbing, the rules, and the people. That’s also where other claims leaders can learn the most from it. Start with a workflow that already runs, review it before automating it, and keep the claims team at the center of the rollout, because the people closest to the work are the ones who make it stick.
