The AI-data platform that sees, reads, and decides with you.
Built exclusively for insurance and owned by your team — ingest anything, automate with editable agents, and query your whole book in one place.
A platform built for insurance — not a wrapper around a model.
Generic automation breaks on the messy reality of submissions. A raw model gives you answers you can’t audit. Inaza is neither.
- Scripts tied to fixed screens and field positions
- One layout change and the workflow snaps
- No understanding of ACORD, FNOL, or SOV
- Answers you can’t trace back to a rule
- No appetite, thresholds, or guidelines applied
- Re-prompting instead of editable logic
- Insurance-native: reads ACORD, FNOL, SOV, loss runs
- Rules auditable and editable 1:1 by your team
- Live in weeks — your experts own the logic
Three moves, one unbroken pipeline.
A submission lands, editable agents work it, and every record is yours to query — no handoffs, no batch jobs, no waiting on us.
- 01Ingest
Meet submissions wherever they arrive.
Email, upload, file storage, an API push, or a portal in the browser — Inaza ingests every file type and turns it into one structured record. Pre-built extraction reads ACORD forms, FNOL emails, SOVs, loss runs, policy declarations, voice notes, and photos.
EmailUploadFile storageAPI pushBrowserPre-built for ACORD · FNOL · SOV · loss runs - 02Automate
Three pre-built agents run the workflow.
Extract pulls structured data from loss runs, ACORD, SOVs and decs. Enrich calls 100+ enrichment APIs — Verisk, LexisNexis, credit bureaus — with no custom integration code. Review acts as an underwriter or adjuster, applying your appetite, thresholds and guidelines. Rules stay editable 1:1 by the people who own them.
ExtractEnrichReview250+ templatesNo coding — your domain experts edit the rules - 03Explore
Every submission, in one queryable warehouse.
Each record is auto-indexed into a single warehouse you can interrogate four ways: Chat for plain-English questions, Dashboards for BI without a data team, Benchmarks against 150+ industry references, and Forecasting with Monte Carlo, ARIMA, and scenario modeling.
ChatDashboardsBenchmarksForecasting150+ benchmarks vs. Aon · Howden · WTW
Pre-built agents that do the work of a desk.
Three production agents ship on day one, backed by a library of templates and an enrichment layer you don’t have to integrate. The logic is yours to read and to change.
- agent templates
- 250+
- enrichment APIs
- 100+
- rules editable by experts
- 1:1
- 01Extractreads the file
Turns loss runs, ACORD forms, SOVs and policy declarations into clean, structured fields — every value traceable to where it came from.
ACORD 125Loss runsSOVDecs - 02Enrichadds the context
Calls 100+ enrichment APIs — Verisk, LexisNexis, credit bureaus — to cross-check and complete the record. No custom integration code to write.
VeriskLexisNexisCredit bureaus - 03Reviewmakes the call
Acts as an underwriter or adjuster: applies your appetite, thresholds and guidelines, and explains the decision against the rule it followed.
AppetiteThresholdsGuidelines
Ask your whole book a question.
Every submission is indexed into one warehouse, and four surfaces sit on top of it. Switch between them — the data underneath never moves.
Plain-English queries
Ask in the language you already use. Inaza resolves it against the live warehouse and shows its working.
The engine room behind the moves.
Four components do the heavy lifting underneath. Each is a page of its own — here’s what they are.
- 01Decoder
The AI engine. Integrates issuance, email, phone and CRMs into a single source of truth, then makes safe, explainable decisions on the data.
Explore - 02Connectors
Pre-built integration engines — IMS/Vertafore, Guidewire, Modotech — ingest, normalize and distribute data across your systems.
Explore - 03AI Data Stream
Data is validated, cross-checked and enriched the moment it enters the system. No batch jobs, no waiting.
Explore - 04Data Warehouse
Cross-organization insight in one queryable store — real-time dashboards and scheduled, pre-built reports out of the box.
Explore
Point checks you can call on their own.
Discrete, production-grade APIs for the focused tasks inside a workflow — each callable directly, no platform commitment required.
- sigSignature ChecksVerify and match signatures across documents.
- dmgVehicle Damage AssessmentAssess damage severity from claim photos.
- altAltered Image DetectionFlag manipulated or doctored images.
- docDocument ParserTurn any document into structured fields.
- cvdCommercial Vehicle DetectionIdentify commercial vehicles in imagery.
- mmlMMLP RecognitionRead make, model, licence and plate.
Built for the team that reads the security questionnaire.
Enterprise controls aren’t a roadmap item. They ship on day one, and we put them in writing.
Model privacy
We never train on your data.
Your submissions and PII never train our models — or anyone’s.Your submissions, documents, and PII are never used to train our models — or anyone else’s. Your data trains your agents, nothing more.
Data minimization
PII redaction on by default.
Sensitive fields are masked before any document reaches a model.Sensitive fields are detected and masked before a document ever reaches a model. You choose what’s visible, down to the field.
Residency
Region pinning — US or EU.
Pin processing and storage to your region — in the contract.Pin processing and storage to your jurisdiction. Data residency you can put in the contract, not just the FAQ.
Access & audit
SSO, scoped roles, full audit trail.
Every action logged and exportable. SAML SSO and scoped roles.Every agent action is logged and exportable. SAML SSO and least-privilege roles map to the policy your security team already runs.
See your own submissions run.
Bring a real file to a working session. We’ll read it, route it, and hand you the logic — live, in weeks not quarters.