Signals are Real, and the Value isn't Where We Expected: Isomer's 1H Progress
When 2026 started, we had a strong conviction and a lot of open questions. The conviction: the story of every insurance claim already exists, scattered across the communications flowing through a claims operation. Emails, faxes, call transcripts, public records. Buried in that flow are early warnings of litigation, missed recovery opportunities, and quiet signs that a routine claim is about to become expensive. The open question: could we surface those stories reliably, at scale, in production?
Six months later, the answer is yes. Here's what we built, what we learned, and why we're even more energized for the second half of this year.
The foundation came first
Before I talk about this year, it's worth being honest about the last one. As a founder, it's the stuff we dream about: how innovation comes together, not in a vacuum, not without Q&A, not without late nights. For us in product management, nothing we shipped in the first half of 2026 was slapped together in six months.
We spent nearly all of 2025 building the foundation, with a relentless focus on one thing: quality. High-quality, auditable, repeatable AI workflows. The ability to run the same AI process a thousand times and trust the results, along with full traceability from every output back to its sources and reasoning.
Our initial focus was to solve one of the most nebulous parts of the claims workflow: the intake, the stuff before FNOL. In building technology to solve the front end of loss, we knew that work would look invisible in a demo. It was tempting to skip ahead to flashier features, but insurance already has enough point solutions, each trying to solve one slice of the workflow. We didn't chase that, because in insurance, an AI system that can't explain itself, can't be audited, and can't produce consistent results isn't a product. It's a liability.
That patience and persistence have paid off. When we committed to claims intelligence this year, we were building on a hardened platform that already knew how to make AI dependable. It's how a small team moved this fast without sacrificing standards.
We found our focus
The most important thing that happened in the first half wasn't a feature. It was a decision.
Early-stage companies are pulled in every direction. For Isomer, in the first half of 2026, we refined the pain point we wanted to solve, leading to the product-market fit: the intelligence and AI enablement layer for carriers.
In practice: we make sense of claims information and all of its surrounding context without losing anything along the way. Every message, document, party, and event is captured, connected, and preserved in full. That completeness is what lets the right information surface right away for claims teams, and what lets carriers accelerate every AI initiative they have, powering AI agents from any vendor, on any surface.
Risk signals were the first and clearest expression of that focus: read a claims organization's communications, detect the risks and opportunities hiding in them, and surface them to the right person before they become losses.
That clarity shaped everything downstream: what we built, who we hired, and how we measured ourselves. The lesson for any product team: depth compounds, breadth dilutes.
From proof-of-concept to production
A year ago, much of what we showed was a vision. Today, our headless platform runs in production, monitoring live claims communications end to end: reading messages and documents as they arrive, matching them to the right claim, and evaluating each one against a large, configurable library of risk signals, from fraud and litigation to recovery and compliance.
Real-world claims communication is messy. Claims get referenced in a dozen inconsistent ways, starting with the foundation: the policyholder showing up with different names and addresses. Making AI dependable in this manner isn't about building a clever model. It's the unglamorous work of data quality, testing discipline, and honest measurement, and where our year on the foundation is proving its worth.
Signals are real, and the value isn't where we expected
What live claims data taught us is that the most consistent value is quieter: the recovery opportunity nobody spotted, the demand letter sitting unread in a crowded inbox, the early signs that a claim is heading toward litigation.
These aren't rare events. They happen every day in every claims operation and add up to real loss, and what the industry politely calls "leakage." Our platform catches them because of its consistency, and the fact that it never gets tired, never skims, never guesses, and never chooses to give the last message of the day the same attention as the first.
This isn't about replacing adjuster judgment. It's about making sure nothing reaches the adjuster's desk unnoticed.
The Isomer Core: The Engine Behind the Layer
The deeper we got with claims organizations, the clearer one thing became: every carrier is pursuing AI right now, through copilots, automation pilots, internal tools, and vendor evaluations. Nearly all of those initiatives hit the same wall: the foundational data underneath isn't ready.
We made a deliberate architectural bet that this intelligence should be open. Through modern integration standards, including MCP (the emerging protocol that lets AI assistants securely access enterprise data) and well-designed APIs, any AI tool a carrier chooses can plug directly into the Isomer Core. Your copilot, your internal chatbot, your analytics platform, your next vendor: all of them get better answers because they're drawing on clean, mapped, always-fresh claims data instead of raw document piles. Be the layer, not the silo.
At Isomer, this is the problem we quietly solved, and it may be the most strategically important thing we built. As our platform processes claims communications, it doesn't just detect risks. It builds a clean, connected picture of the entire claims operation: every claim, party, document, and event, mapped to industry standards and kept continuously current. We call it Isomer Core.
The team
None of this happens without the people. We're a small team that shipped meaningful products nearly every week while raising the bar on quality, security, and enterprise readiness. Watching this group has been the best part of my year: engineers debating claims workflows with genuine curiosity, going deep with adjusters and claims executives to understand the real job.
We can’t wait to share what’s on deck for the second half of 2026. Follow us here.