4 MINUTE READ
What does it take to build a modern claims platform that supports P&C insurance operations globally?
In this episode of Policy in Wonderland podcast, Michael Krikheli, co-founder and CTO of Five Sigma, sits down with host Evyatar Amira to share the behind-the-scenes story of how the company got started, the challenges of building an AI-native claims software before AI was trending, and how that journey led to the creation of Clive™, the industry’s first multi-agent AI claims expert.
From the early days of MVPs to shaping real automation for adjusters, this conversation dives into the thinking, process, and vision that brought Five Sigma to where it is today.
How Five Sigma Started Building AI for Claims Management
Q: Hey, welcome to another episode of Policy in Wonderland. Today we’re diving into a topic that’s super interesting to me – both from my background as a lawyer and an underwriter – the world of insurance claims. But not just claims, we’ll explore the space through someone who’s a veteran in the insurtech world. It’s going to be fascinating to hear about the past, present, and the tech-driven future of claims.
Without further ado, I have the pleasure of welcoming Michael Krikheli, Co-Founder and CTO of Five Sigma. Michael, how are you?
A: Great, thanks for having me! Excited to be here.
Q: Awesome. So, let’s get the basics out of the way – what is Five Sigma, and how did it all start?
A: My background is in tech and applied mathematics. After my military service, I met my co-founders and we started exploring what to build.
Not all of us came from insurance. Some of us were tech-minded folks looking for a real problem to solve.
Q: Okay, and how did the name Five Sigma come about?
A:
Well, we were five founders – that’s the “Five.” And “Sigma” comes from physics – it’s a term that represents high certainty. We wanted something that signaled precision. It was actually picked before we fully decided to focus on insurance.
Q: So how did insurance come into the picture?
A: We were looking for meaningful, painful problems in big industries. We talked to healthcare, cybersecurity, and then insurance companies. At first, we didn’t even know what we were looking for – we just asked smart people in insurance: “Do you have a data problem we can solve?”
And some of them said: “Yes, claims are a mess.”
They mentioned things like “jumper claims” – small claims that suddenly escalate into large ones. Or problems understanding documents. So we got intrigued and started digging into it.
Q: You were early into this, right?
A: Yeah, this was 2018. We built our MVP and sold our first systems that same year.
Q: What was the core insight?
A: We realized that current claims systems don’t tell the full story. A lot of the data isn’t in the system – it’s in Outlook, phones, shared folders. You can’t just “run AI” on top of bad or incomplete data. So we said: let’s build a modern claims system where the data structure reflects the actual process.
From there, you can layer automation and AI. For example, once everything’s recorded automatically, you can trigger workflows or surface insights.
Q: How did insurers react? Especially back then when AI wasn’t as mainstream.
A: At first, we tried to say “we’ll give you AI,” but quickly learned that wasn’t enough. So we shifted: we’ll give you the entire modern claims system, fully SaaS, fully cloud-native, easy to configure and deploy.
And originally, we even thought about being a TPA ourselves – doing the claims handling too. But customers didn’t need a TPA, they needed great software.
Q: So you’re not a TPA or MGA?
A: Exactly. We’re a pure SaaS claims platform, focused on Property & Casualty (P&C) lines. And yes, in 2018, cloud in insurance wasn’t a thing. SaaS wasn’t even in the conversation yet.
So when we came in and said: “This won’t take years to implement, won’t cost millions, and your adjusters will learn it in an hour,” that really clicked.
Our first go-to-market was targeting newer MGAs and carriers without legacy systems.
Q: And what about traditional insurers?
A: Even at traditional carriers, the claims execs are often the same people who worked at newer insurtechs. So they spoke our language. Once we built credibility, it got easier to scale into larger accounts.
Five Sigma’s Automation-First SaaS Claims Management Technology
Q: How does the product work today?
A: Our core platform handles the entire claims process: FNOL, coverage decision, fraud checks, liability, communication, payments, task and workflow management, dashboards, and reporting. Everything.
Our philosophy is: Let automation do 90% of the work on 100% of claims. The human adjuster steps in only where judgment or empathy is needed.
Q: How did COVID impact you?
A: Wild story. I was in NYC visiting a prospect. The weekend hit and suddenly everyone was talking about this virus. Stores were out of masks. Flights were getting canceled. I booked a one-way ticket back and made it home just before lockdown.
Business-wise? Everything froze for a minute. But then – boom. Everyone went digital. Zoom took off. People realized they don’t need to travel for every meeting.
It actually helped us – insurers saw they needed modern, remote-capable tools.
Q: What changes stuck from that era?
A:
Remote work. Remote inspections. Handling claims via video, photo, text. Adjusters no longer needed to sit in offices and use old desktops. Our cloud-native system worked from a laptop and headphones.
Q: What about communication?
A: That’s key. We built all communications into the claims platform – email, text, calls, WhatsApp. You don’t need a phone, everything is handled from within the platform. Everything is indexed, recorded for compliance, and searchable, so adjusters can see context, reduce errors, and make decisions faster.
Using AI to Turn Claims Data Into Actionable Insights
Q: How do you see the market today?
A: Claims are rich with data. Unlike policy issuance, they’re interaction-heavy, messy, and complex. Every claim is different. But now with GenAI, you can summarize, analyze, and act on unstructured data much more easily.
The challenge is you can’t just throw ChatGPT at a 10 year-old bodily injury claim. You need proper workflows, controls, audit trails, and fail-safes.
Q: So what makes Five Sigma different?
A: We’re focused. That’s our edge. We built a model for how claims really work. We understand the real-world messiness of workflows and data. That’s hard to replicate.
In 2024, we introduced Clive™, the insurance industry’s first AI Multi-Agent Claims Expert, that works on top of any claims management system (CMS) – even old legacy systems.
Q: Why can’t insurers just build this themselves?
A: Some try. They’ve got great engineers. But AI in claims is not just about the tech – it’s about understanding processes, edge cases, compliance, and real-world delivery. Many projects get stuck in demo mode and never go live.
Q: What’s still the “Holy Grail” in claims tech?
A: We call it “Automation First”. Not 100% automation on 20% of claims — but 90% automation on 100% of claims. Even big, complex claims still benefit from partial automation: summarizing documents, suggesting next steps, flagging fraud. AI will handle most of the claim, and humans will add a human touch.
That’s what we’re enabling — and that’s what the future looks like.
Q: I like that! Thank you for your time, Michael.
A: Great! Thank you for having me.
