# Bevaya overview

Insurance carriers receive a constant stream of high-stakes documents — claims, submissions, loss runs, endorsement requests, invoices, legal demands — and the cost of processing them slowly, incorrectly, or inconsistently compounds across every downstream operation. The teams handling this work aren't struggling because they lack effort or expertise. They're struggling because the tools they use extract information but don't operationalize it: documents get read, but the work of determining what they mean, where they belong, and what should happen next still falls on people.

Bevaya closes that gap. It is controlled AI automation for document-heavy insurance operations — an operating system for the seam between inbound information and trusted system-of-record action. Pre-built AI Agents cover the most common carrier workflows out of the box and can be configured and deployed in weeks, not months.

## What Bevaya does

Bevaya is workflow-first, not document-first. It doesn't stop at reading a document. For every inbound case, Bevaya:

1. **Classifies** the document type and workflow intent.
2. **Extracts** the fields that matter, with field-level confidence.
3. **Matches** the document to the correct claim, policy, submission, audit record, or intake path.
4. **Decides** the next operational step based on your approved business rules.
5. **Acts** — either writing back to your system of record automatically, or routing a prepared exception to a reviewer with the evidence and recommendation they need to resolve it in one touch.

Routine, low-risk cases flow through without human handling. Ambiguous, high-risk, or system-failed cases surface to reviewers with full context already assembled. The result: fewer manual handoffs, faster cycle time, and a complete audit trail from source document to action taken.

## What Bevaya automates

Bevaya automates the document-heavy workflows that sit across underwriting, claims, and policy servicing. Each automation runs as an AI Agent that follows your defined business process.

**Underwriting** — submissions land, get classified and extracted, and flow into your underwriting system clean and ready for a decision. Loss runs, schedules, ACORD forms, and renewals follow the same pattern: structured, matched, and routed before they reach the underwriter's desk.

- Submission Intake
- Loss Run Processing
- Schedules & SOV Extraction
- ACORD Form Extraction
- Policy Renewal Handling

**Claims** — documents arrive, get indexed and summarized, and reach the right adjuster with legal demands and medical bills already prepped. Claims are created or updated in your system of record automatically; only the uncertain cases surface for human review.

- Claims Setup
- Claim-to-Policy Comparison
- Claim Indexing
- Legal Demands Identification & Extraction
- Medical Bill Identification & Extraction
- Invoice Payment Processing
- Claim File Summarization

**Policy servicing** — endorsements, certificates of insurance, and premium audits move from inbox to system without re-keying, at the pace your brokers and policyholders expect.

- Endorsement Processing
- COI Creation
- Premium Audit Processing

Each pattern is documented in [Use Cases](../use-cases/overview.md) with the business problem, trigger and source data, AI Agent behavior, human review model, required integrations, and a recommended rollout path.

## Who uses Bevaya

Bevaya is built for P&C carriers, brokers, and TPAs. Within those organizations, several roles work with the platform:

**Operations and transformation leaders** — claims operations, shared services, intake and triage, and automation teams — own the outcomes Bevaya improves: straight-through processing rate, intake speed, backlog size, SLA attainment, and resilience during catastrophe surges. They evaluate Bevaya against operational and financial KPIs, not feature counts.

**Implementation teams and platform administrators** set up the organization, create workspaces and AI Agents, manage users and access, and configure the environments where AI Agents run.

**Builders** create, configure, test, and publish AI Agents. They use the workflow builder to compose flows from triggers and nodes — defining how each inbound case is classified, extracted, matched, routed, reviewed, and written back.

**Reviewers** are the operational specialists who handle the exceptions Bevaya escalates for human judgment. Instead of starting from a blank form, they open a prepared workspace: the raw document, a proposed classification, extracted fields, a recommended match, and a suggested action. Their job is to validate or correct — not reconstruct.

**Developers and integration engineers** connect Bevaya to carrier systems of record through the API and integration nodes, and automate runtime operations such as triggering runs, retrieving execution details, and inspecting item status.

## How Bevaya works

The platform follows a lifecycle that most teams move through in sequence, then operate in a continuous loop.

1. **Onboard.** Set up your organization, workspaces, and AI Agents, then invite users and assign access. See [Onboarding & User Management](../onboarding/platform-hierarchy.md).
2. **Build AI Agents.** Create an AI Agent from a use-case template or from scratch. Define its business process by composing nodes into a flow: a trigger ingests the input, then nodes classify, extract, match, route, send exceptions to human review, and write back to your system of record. Test the draft against real documents, then publish. See [Build AI Agents](../build-ai-agents/overview.md).
3. **Monitor & review.** Once you publish and activate the flow, its configured trigger starts production workflow runs (publishing alone doesn't start processing — you activate the flow to put its trigger live). Browse run history, inspect node-level execution details, work the exceptions routed to human review, and track item status across Failed, In Progress, Review, and Complete. See [Monitor & Review](../monitor-review/overview.md).

If you are new to the platform's vocabulary — organization, workspace, AI Agent, flow, trigger, node, workflow run, work item, and human review — start with [Platform concepts](./platform-concepts.md).

## Where to go next

- **New to Bevaya?** Read [Platform concepts](./platform-concepts.md) to understand the core model, then browse [Use Cases](../use-cases/overview.md) to find the workflow you want to automate.
- **Setting up the platform?** Start with [Onboarding & User Management](../onboarding/platform-hierarchy.md).
- **Ready to build?** Go to [Build AI Agents](../build-ai-agents/overview.md).
- **Integrating through the API?** See the [API overview](/api).