Managed cloud
For agencies that prefer a managed deployment. Before deployment we disclose the hosting regions, data residency, model providers and processing arrangements that apply to you.
Deployment options
Your AI. Your infrastructure. Your control.
Insurance organizations have different requirements for data residency, infrastructure ownership, model hosting and operational security. Bexara AI's architecture is designed to support different operating environments without changing the core product.
Architecture
See where processing happens.
Select a deployment model. The sample document shows the path through the components, and the labeled boundary shows what stays inside your control.
Agency users
Carrier websites
Agency management system
Identity & access
Validation & approval
Audit logs
Integration endpoints
Bexara orchestration
AI processing
Data storage
Bexara AI operations
External model provider
Shared responsibilities
Compare
What runs where.
| Managed cloud | Private cloud | On-premises | |
|---|---|---|---|
| Application & execution services | Bexara AI | Your cloud account | Your data center |
| Infrastructure operations | Bexara AI | You (with Bexara tooling) | You |
| Model execution | Disclosed per agreement | Private endpoints or self-hosted, where supported | Locally hosted, where supported |
| Document storage | Managed, tenant-scoped | Your storage and keys | Local storage |
| Identity | Your IdP via SSO, where implemented | Your IdP | Your directory services |
| Software updates | Managed by Bexara AI | Agreed release windows | Your approved process |
| Carrier portal access | Outbound from managed cloud | Outbound from your network | Outbound from your network — still required |
| External model providers | Only if configured | Only connections you configure | Not used with local models |
Private cloud
For organizations that need more infrastructure control. Responsibilities are shared: you own the account, network and keys; Bexara AI provides the platform, deployment tooling, updates and support.
On-premises
For organizations that need supported AI processing inside their own environment. Workflows that reach carrier portals or cloud systems still need those outbound connections.
Disconnected (air-gapped) environments. If a disconnected configuration is offered, it is documented separately with its limitations. For example, carrier portal retrieval and cloud-hosted AMS integrations aren't possible without outbound access, and updates must be delivered through an offline process.
Availability. Each deployment model is offered only where the implementation, operating procedures and support arrangements are in place. We confirm what applies to your organization during scoping.
Choose the deployment architecture that fits your operational and security requirements.
Tell us about your infrastructure, data residency and model hosting requirements, and we'll walk through the options that apply.
Each demonstration is shaped around your agency's workflows, systems and security requirements.