Security built for insurance operations
Secure, accountable AI agents for claims and servicing. Arvy protects policyholder data, and your team decides what agents can access, what they can do and when a person takes over.
Our SOC 2 Type II audit is in progress. Follow it in the Trust Centre.
Enterprise-grade security
Every Arvy deployment is set up to protect your data: encrypted, isolated from other customers and open only to the people and systems you approve.

Encrypted in transit and at rest
Data is encrypted on its way to Arvy and wherever it is stored.
Isolated environments
Your deployment runs in its own environment, separate from other customers’ data.
Role-based access
SAML and OIDC single sign-on with SCIM provisioning, and roles scoped down to a single workflow.
PII redaction
Card numbers, IDs and other sensitive details are masked before an agent sees them.
Backups and recovery
Automated backups, with recovery targets agreed for your deployment.
Requirements agreed up front
We review data access, processing and retention requirements with your IT and security teams before launch.
Your data, your rules
You choose where policyholder data lives, how long it is kept and what AI models may do with it.
Data residency
Choose Canada or US hosting for your deployment.
Private cloud deployment
Run Arvy in your own cloud environment when data has to stay inside your boundary.
No training on your data
Your interactions are never used to train shared models. We commit to that in writing.
Zero data retention
Model providers can be held to zero-retention terms, so prompts and responses are not kept after processing.
Retention you set
Conversation records are kept for the period you choose and deleted on request.
Your systems of record
Arvy completes work inside the policy, claims and agency systems you already use.
Every action on the record
Arvy agents follow the workflows your team approves, and each step is logged. You can review what an agent did and when a person took over.
Approved workflows
Agents act only within the information sources, actions and identity checks your team approves.
Guardrails
Policy checks on every reply, with blocked answers routed to a person.
Human handoff
Your team steps in when a case needs judgment, with the full context in front of them.
Decisions stay with you
Coverage decisions, policy changes and claims judgments remain with your people.
Tested before launch
We test each workflow and its handoffs with your team before it takes live requests.
Audit log
Every agent action and handoff, timestamped and ready for review.
Ready for your security review
Send us your questionnaire and we’ll complete it with your team. The Trust Centre holds the documents security and procurement reviewers ask for most.
Trust Centre
Policies, controls and SOC 2 audit progress in one place.
Security questionnaires
Send your vendor questionnaire and we’ll complete it with your team.
Architecture and data flows
What Arvy connects to, what data it handles and where that data goes.
Subprocessors
The vendors that process customer data, and what each one does.
Audit timeline
Where our SOC 2 Type II audit stands and when to expect the report.
Scoped pilots
Start with one workflow and approved system access, then expand.
Where is our data hosted?
Is our data used to train AI models?
What can an Arvy agent access and do?
Does Arvy make coverage or claims decisions?
How do we start a security review?

