The insurance & risk vertical
Decision automation and AI systems that can't hallucinate their way into a claims file — deterministic guardrails, human sign-off gates, and audit trails built for the NAIC AI-bulletin era. Engineered by people who have shipped into the world's largest carriers and passed their scrutiny.
The pedigree, in specifics
Our founder co-architected Sapiens Decision — the enterprise decision-management platform born at Freddie Mac and grown into the insurance-software major's flagship — adopted by some of the world's largest financial institutions.
Underwriting rules, rating logic, eligibility, policy decisions — captured as governed, versioned, testable models that business people can read and auditors can trace. Not code archaeology.
When every automated decision must be explainable — to an auditor, a regulator, a court — the decision layer needs provenance built in. We've architected exactly that, at the highest level of the industry.
Where we're deep
Multimodal accident evaluation — crash data, imagery and visual simulation turned into damage and risk assessments a carrier can defend. Built it, shipped it into Fortune 100 carriers, passed the audit that came with it.
High-volume vehicle telemetry — sensor streams, event data, video — ingested and engineered at carrier scale, with the data-handling controls Fortune 100 security teams demand.
Policy, claims and insurance-operations platforms — from first notice of loss to resolution, every workflow traceable, reviewable and compliant.
Claims decisioning that moves faster without going rogue — deterministic rules where it matters, governed AI where it helps, audit trail everywhere.
How we govern AI
Platform-grade AI governance, delivered as engineering — specific, checkable commitments, not a values page.
What runs in production is deterministic code — never a model improvising. An AI system cannot invent a claim decision, a rate, or a coverage answer, because no AI is in the decision path unless you put it there deliberately.
Where runtime AI does earn a place — a chatbot, a triage assistant — it operates inside guardrails: scoped knowledge, consistent behavior, and human checkpoints on anything consequential.
Every decision, every change, every access — logged with provenance. Documentation your compliance team can hand to a regulator under the NAIC AI Model Bulletin, now adopted across much of the country.
Client code and data are never used to train anything, and never leave your environment. Your security team can verify it — we'll show them where to look.
Proof in production
We took a risk-intelligence startup with no internal engineers to SOC 2 Type II and live deployments with Fortune 100 carriers — as their entire engineering department.
Read the outcome →A short readiness conversation — in your language, about your product and the carriers you're chasing.
Talk to a team that's done this in insurance