2 hours ago
Joel Raedeke outlines a client-first model for transformation in casualty claims
Transformation
By Chris Davis
Technology transformation in casualty claims does not begin with artificial intelligence. For Joel Raedeke (pictured), US chief technology officer at Crawford & Company, it begins with the client. “The North Star for all things technology, AI, and analytics is the client,” Raedeke said. “The question is: what does success look like, and what is valuable to the client?”
That principle shapes how Broadspire evaluates innovation across casualty claims - from first notice of loss through closure – and how the company structures the teams responsible for delivering it.
Four ways clients define value
Across corporate risk managers, carriers, MGAs and MGUs, Raedeke said value in claims operations generally falls into four categories. The first is experience. “The first concept of value is ‘take care of my people,’” he said. “It is an experience-based concept.” Adjusters often function as extensions of a client’s internal team, while claimants and risk management staff also fall into that “my people” category.
A second dimension focuses on efficiency. “A client might say, ‘I want to pay as little as possible for the process,’ so efficiency is a type of measure,” Raedeke said. That focus appears more frequently among carriers and MGUs, while corporate clients may emphasize the claimant and employee experience more heavily.
Quality is a third measure of value. “Especially when a chief claims officer is engaging, they can say, ‘For me, value is SLAs [service level agreements],’” Raedeke said. “I want to make sure that whatever process you are running meets certain key quality criteria.”
The fourth dimension is outcomes. “Whether you are a corporate, carrier, MGA, or MGU, it is about the loss cost or loss pick that you are paying,” he said. The economics of claims make this category particularly important. “Often it is a 10-to-1 ratio,” Raedeke said. “For every million dollars in fees, you might have $10 million of loss costs.” Analytics and AI that influence claim outcomes can therefore generate the most significant financial impact.
Turning complexity into opportunity zones
Casualty claims operations span a wide landscape – from first notice of loss through resolution and across multiple lines including auto liability, general liability and workers’ compensation. “The terrain that these claim processes occur in – from day one, first notice of loss, all the way to closure – covers a lot of ground,” Raedeke said.
To manage that complexity, Broadspire breaks the environment into smaller domains that Raedeke describes as “opportunity zones.” “We try to make each zone small enough to be manageable, but large enough to have rich veins of customer value,” he said.
Each zone is assigned a cross-functional team with the expertise required to identify opportunities and develop solutions. These teams can include subject matter experts, data scientists, analysts, UX designers, engineers and architects. Operational specialists such as adjusters or nurses may also be embedded when the work relates directly to claims handling or medical management.
“We set them up almost like a small business,” Raedeke said. “They are given the boundaries of the terrain where they focus and they have all the expertise they need and all the authority they need to create value in that terrain.”
The model mirrors structures adopted across the technology sector. Amazon refers to “two-pizza teams,” Gartner calls them “fusion teams,” and Spotify uses the term “squads.” Regardless of the label, the goal is the same: empower small teams to move quickly within a defined domain. “Their mission is to find and harvest the opportunity,” Raedeke said.
Measuring results like an investment portfolio
Autonomy does not eliminate accountability. Teams report regularly on measurable results tied to their domains, typically on a monthly cadence. Those metrics may include experience indicators such as CSAT [ customer satisfaction score] and NPS [net promoter score], along with operational and outcome measures tied to claim costs.
In some cases, teams conduct control studies to measure the effect of a newly deployed tool or model. “They are doing control studies that demonstrate the solution they just invented and deployed is, in a control-tested way, delivering a specific benefit on the loss pick,” Raedeke said.
Progress is often summarized in structured updates that resemble investor reports. “Sometimes that takes the form of a 10-page PowerPoint deck that looks almost like an investment portfolio,” he said. The presentation outlines past achievements, current impact and future potential alongside the resources required to continue development. “It says, ‘Here is what we have done in the past. Here is the value we have created to date. Here is the opportunity to create value we see ahead of us.’”
Those reports allow leadership to evaluate initiatives and allocate funding accordingly. “The work has to prove itself as something that aligns to the client,” Raedeke said.
Building AI capabilities from the foundation
When it comes to artificial intelligence, Raedeke said Broadspire generally prefers to build on foundational capabilities rather than rely heavily on specialized vendors for niche solutions. Some insurers take the opposite approach. “Some carriers basically find a different vendor for every target – one vendor for fraud detection, one for subrogation detection, another for something else,” he said.
Broadspire’s engineering teams tend to take a more hands-on approach. “My natural inclination is a high-ownership mindset,” Raedeke said. “My engineering team is the same way. We like to get our feet wet and understand what we are dealing with.”
For example, the company uses platforms such as AWS Bedrock to integrate large language models into workflows that process unstructured and structured data within given workflow scenarios. This allows Broadspire to own the mechanics, the efficacy testing, the usability and the outcomes. “We are not trusting a vendor with that as a black box. This allows us to demonstrate all of these aspects of the value chain to ourselves and our clients.”
When strategic priorities reshape the roadmap
Even with autonomous teams, external forces can reshape priorities. Regulatory developments are one example. “NYDFS [New York Department of Financial Services] comes to mind, among others,” Raedeke said. “Something is coming from the outside and puts new demands on us, whether from a security or privacy perspective.” Such changes can affect multiple software teams simultaneously and require adjustments to development roadmaps.
Major client opportunities can also influence priorities. “A large prospect may want something that would not naturally or organically be on the roadmap,” he said. “But their need can still impose itself.”
Despite those pressures, Raedeke said the underlying philosophy remains consistent: align innovation with measurable client value and empower teams closest to the work to deliver it. “Transformation is not really coming from me,” he said. “I might be the one who sets up the environment, but I am not the one inventing the ideas – they are.”
