Scout InsurTech Rising with Crosstie

Sean Eldridge, Co-Founder and CEO at Crosstie, was interviewed by Andrew Daniels to discuss Crosstie's six and a half years of staying focused on claims and why true end-to-end workflow orchestration is the industry's next hard problem. Crosstie's platform is embedded directly into claims and operational workflows for carriers, TPAs, and self-insured organizations.

Sean, the insurtech space is crowded with platforms trying to do everything. What made you commit early on to claims, and only claims, and what has that singular focus unlocked that a broader approach couldn't?

“Most people in claims will tell you they fell backwards into it. My co-founder John and I jumped in headfirst.

We were building a different business when we came across the Workers' Compensation Benchmarking Study and fell in love with the problem. The claims process seemed overly difficult for everyone involved, which felt like a problem worth solving. So we sunset that previous business and went all-in on what would become Crosstie, which we built out of Harvard’s Innovation Lab.

Before writing a line of code, we spent roughly 5,000 hours speaking with claimants, adjusters, claims leaders, risk professionals, attorneys, vendors, and other stakeholders. That research gave us a solid understanding of how claims work and helped us see where the friction lived.

I'm not sure we would have been successful if we had tried to take on other functions such as underwriting, other segments such as life and health, or even broader service workflows from day one. Claims gave us the foundation. For our first six years, we stayed focused there: claimant communication, document processing, intake, voice automation, claim intelligence, and the operational handoffs that move claims forward.

Only more recently have we expanded into adjacent service and loss control workflows, and even that expansion came directly from what we learned in claims. The same capabilities that improve claim outcomes, including multimodal communications and workflow automation, can help insurers better serve policyholders and other parties, including agents, brokers, attorneys, and vendors, while also collecting better insights for risk engineering and ultimately underwriting.

Six and a half years of staying focused and listening carefully to the people actually doing this work has taught us more than we ever could have figured out on our own.”

Six and a half years is a long time in insurtech. How has building through multiple market cycles and directly from customer feedback shaped your product in ways that newer entrants simply can't replicate?

“When we launched with our first customer in 2020, we could count on one hand the number of other insurtechs in our competitive landscape. Two years later, that number was closer to fifteen. Four years later, it had doubled again. Now, six and a half years in, the number of competitors has doubled again.

That's been interesting to watch. In 2020, we were a bit of an oddity because we didn't come from P&C insurance. Today, a lot of new entrants are coming from the technology side, drawn to the size, complexity, and legacy infrastructure of the industry. That can be a strength, but it also means many teams are learning the operating reality of insurance while they build.

What has stayed consistent is that insurance buyers care whether the product works inside their operation, integrates with their systems, and doesn't create another burden for their teams.

Because we've been in production since 2020, our roadmap has been shaped inside live claims operations. Customers pulled us from claimant communication into document workflows, from documents into voice and broader automation, from automation into intelligence, and from intelligence into orchestration. We didn't plan that path. We listened closely to where our customers were experiencing friction.

That has created a head start of several years that is hard to compress. Our technology is being used by some of the highest-volume tier 1 carriers and multiple top-10 TPAs, which has given us a rich view into the edge cases, integration constraints, and operational handoffs that determine whether a product actually works at scale. Competitors can move quickly, and we assume they will. But we’re building on years of production learning, customer trust, and domain pattern recognition that take time to replicate.”

There's a lot of AI conversation focused on adjuster efficiency, but you've talked about balancing that with the policyholder and claimant experience. Why does that balance matter, and where do you see most platforms getting it wrong?

“Adjuster efficiency and claimant experience are often the same problem viewed from different sides.

When claimants are confused, they call more. When they don't know what happens next, they follow up more. When forms are hard to complete, documents come in late. When communication feels inconsistent, frustration builds and attorneys get involved. All of that creates more work for adjusters.

That's why internal efficiency and external experience need to be designed together. A tool that only helps the adjuster may still leave the claimant in the dark. A tool that only improves the front-end experience may still leave the adjuster with the same manual work behind the scenes. Many solutions get this wrong by treating customer experience and operational efficiency as separate categories, when in claims they are tightly connected.

The claim is the promise of insurance. It's the moment the entire relationship between a carrier and a policyholder gets tested. AI should make that moment better for everyone involved, not just faster for the operation. The best use of AI is to remove the administrative drag around the claim so professionals have more time for the work where judgment, context, and empathy matter most.”

Carriers and TPAs often say yes quickly but then take 6 to 18 months to actually implement. What does that dynamic tell us about where the real friction in claims modernization lives, and how should vendors be thinking about it?

“It tells us that buying a solution and deploying it are two different conversations inside the same company.

Most claims leaders already know where the pain is. The challenge is that saying yes to a new technology is only the beginning. Claims leadership then has to marshal IT resources to enable the integration work, and IT has its own queue and its own priorities. Most buyers want full integration. They don't want their adjusters managing another screen. That's the right instinct, but it means the implementation can be more involved than a standalone deployment.

The way we've tried to address this is by meeting customers where their data already lives. We've built integrations across major claim systems and developed our own insurance-native ETL so we can exchange data without requiring customers to match our specifications. In many cases they can repurpose file feeds and connections they've already built, and we handle the mapping work on our end.

The vendors that implement fastest will be the ones that reduce lift on the customer's side. The ones that leave buyers to figure it out on their own are the ones that end up with frustrated champions trying to explain why the solution they bought still isn't live.”

Claims has come a long way, but what are the problems that still don't have good answers? Where is the industry still flying blind, and what needs to happen from carriers, vendors, or regulators before those gaps get closed?

“The problem that doesn't have a good answer yet is true end-to-end workflow orchestration, and I think the industry is only beginning to understand how hard that actually is.

Over the last few years, claims organizations have invested in point solutions to solve their most pressing problems. Document processing here, agentic voice there, a claimant portal somewhere else. Those investments made sense, and many of them delivered real value. But now those same organizations may be looking at a collection of vendors that don't talk to one another and realizing that getting to true workflow orchestration means figuring out how to connect them. That's a form of technology debt the industry will need to solve.

Core systems built partner ecosystems rather than building every capability themselves. Point solution vendors specialized deeply in one area because that's where they could win. Both strategies made sense, but neither was designed with end-to-end orchestration in mind.

If you believe the future of claims is agentic, the harder challenge isn't deploying agents to handle individual tasks. It's connecting those agents across the full claims lifecycle so they can hand off to one another, share context, and give claims teams one place to manage the work. That orchestration layer is where we believe differentiated claims outcomes will come from, and closing that gap is the most important thing the industry needs to figure out right now.”