Core-system agnostic deployment runs inside the insurer's own Azure tenant, so claim data never has to leave the environment the insurer already governs.

Isomer™, which reviews every claim communication at the point of receipt, today announced availability on the Microsoft Azure Marketplace as an Azure Managed Application. Insurers can now deploy the full Isomer solution inside their own Azure tenant as a dedicated, single-tenant environment, with data persistence and core processing running entirely within their own subscription. Isomer returns cited risk signals before the claim system sees the file, all within a sovereign deployment: using the insurer's cloud, the insurer's keys, and Isomer's detection inside the boundary the insurer already controls.

"Carriers have been clear that AI adoption depends on control: where the system runs, who holds the keys, and whether every result can be explained," said Sean Chou, founder and CEO of Isomer. "Running Isomer inside the insurer's own Azure tenant answers all three, and the Marketplace makes that available to any insurer on Azure. Where it runs changes. What we built underneath doesn't."

Most AI solutions across insurance ask the carrier to send claim data out. Isomer deploys the other way. The solution installs into the insurer's tenant through an automated Microsoft Azure Marketplace offer, receives upgrades through that same offer, fits the network segmentation and firewall inspection that enterprise security teams already enforce, and can operate without public internet exposure. Insurers are buying a packaged product here, not commissioning a project: there is no integration build, no implementation calendar, and no vendor project team to manage. Customers can bring their own encryption keys, route logs to their own security monitoring, and hold their data under their own controls, and Isomer never trains models on customer data.

The solution is model-flexible by design: insurers can run models through their own Azure AI services and existing model API keys inside the tenant, and open-weight models can run entirely within the environment where data residency demands it. That is data sovereignty in its strictest form: no data or model activity ever leaves a boundary the insurer controls end to end. This architecture was shaped in security and network reviews with some of the largest claims organizations in the world. The detection library, the insurance-specific training, and every cited signal remain Isomer's. The Marketplace listing changes how the solution is delivered, not what built it.

The timing matters because the regulatory ground has shifted. For many review boards, that shift is exactly what has kept AI proposals on hold. Isomer was built to give them something they can approve. The NAIC's model bulletin on insurers' use of artificial intelligence, now adopted in 25-plus states, holds insurers accountable for the AI systems they buy, not just the ones they build. That accountability requires governance, documentation, and the ability to explain outcomes. Isomer was built for that standard: every risk signal the solution raises is cited to its source document and page, from a library of more than 90 detectors and growing. Surveys of claims professionals repeatedly find that lack of explainability is the leading reason they ultimately end up overriding AI. The same audit trail that satisfies an examiner is the one that lets an adjuster trust the signal.

Isomer runs in production today with commercial P&C carriers, MGAs, and third-party administrators. New customers typically see cited risk signals on their own claims within days of going live, not after a multi-quarter tuning period. In current production, adjusters open a structured file rather than an inbox, and risk signals reach the business within a median of 30 minutes from receipt. That earlier signal is what preserves negotiating leverage on time-limited demands and coverage position before litigation posture hardens.

About Isomer

Isomer puts AI into production for insurance carriers, MGAs, and TPAs. It reads every claim communication at the point of receipt, detects risk, and writes structured results back to the claim system, with every signal cited to its source. Isomer was co-founded by CEO Sean Chou, founding CTO of Fieldglass (acquired by SAP), and CPO Sean Casey, and is backed by Equal Ventures and Emergence Capital. Learn more at isomer.ai.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260929130657/en/

Carriers have been clear that AI adoption depends on control: where the system runs, who holds the keys, and whether every result can be explained. Running Isomer inside the insurer's own Azure tenant answers all three.