What does ClaimVision demand from your architecture?

Very little: ClaimVision is an AI claims platform for IT leaders that adds a reading layer to the stack you already run, with no data migration and no new system of record.

It reads the claim documents your systems already hold, returns findings with the evidence attached, and changes nothing about how your core systems store, own or serve the claim. No re-platforming, no big-bang cutover on the system your business runs on.

app.claimvision.ai/upload
The ClaimVision upload view, where the loss reports, policies, FNOL and inspection PDFs this AI claims platform for IT leaders reads go in.
The upload view takes the loss report, the policy, the FNOL and the inspection report as PDFs, the documents ClaimVision reads.

Security and architecture

The facts your security review will ask for

Hosting and tenancy

Enterprise-grade Microsoft Azure environment. Carrier data is logically isolated per customer, and a carrier's claim data never trains another carrier's models.

Encryption and access

Encrypted in transit and at rest. Users see only their own company's claims. No public database port; engineers reach production only through Azure Bastion, with approval-gated read access.

Data flow

Documents in, findings out. ClaimVision holds working copies for processing and the decision record for audit; your system of record stays authoritative for the claim. More on the AI layer on a core claims system.

Auditability

Every AI recommendation, source citation, confidence level and human override is preserved, append-only. It is the first thing compliance and model-governance teams ask about. See the AI claims audit trail.

Rollout shape

Weeks, not quarters: a closed-file benchmark first, with no workflow change at all, then a live parallel test alongside the human process, then a production scope you control. Each stage is independently reversible.

Documentation

Detailed security architecture and data-processing agreements are available under NDA. The public summary lives on the security page.

Where your data lives: Microsoft Azure regions in the United States. The contracting entity, Decision Agency OÜ, is registered in Estonia, in the EU, so GDPR applies to us. SOC 2: ClaimVision is built to SOC 2 principles and we are working towards a SOC 2 Type 2 audit; we are not yet certified, and we'll say so plainly until we are. Security and data handling

What you won't find here: our pipeline internals, engine choices or extraction mechanics. Partly moat, partly principle: a vendor that explains its machinery on a marketing page will explain yours on someone else's.

FAQ

Questions IT leaders ask

Where is ClaimVision hosted?

In an enterprise-grade Microsoft Azure environment, in US regions. Carrier data is logically isolated per customer and encrypted in transit and at rest, and each user sees only their own company's claims.

Does an AI claims platform need access to our core claims system?

Not to start. The first step works on a set of closed claim files, with no change to your workflow. How documents reach ClaimVision after that is scoped with your team, and your core system stays the system of record.

Is our claims data used to train AI models?

No other carrier's models. A carrier's claim data never trains another carrier's models; each carrier's models learn from its own data, and each carrier's data is logically isolated.

What documentation is available for a security review?

Detailed security architecture and data-processing agreements are available under NDA, and the public summary is on the security page. Send your questionnaire early; we would rather clear it alongside the pilot than after it.

Send us your security questionnaire early.

We'd rather clear your review alongside the pilot than after it. Bring your architecture lead to the first call.

Related: AI layer on a core claims system · AI property claims review · ClaimVision and Guidewire ClaimCenter · AI claims audit trail