# Platform concepts

This page introduces the core concepts you need to navigate the rest of the documentation. The concepts fall into three groups: the **hierarchy** that organizes your resources, the **build model** you use to create automations, and the **runtime model** that describes what happens when automations run. Definitions here are intentionally brief — each links to the guide or reference page that covers it in full.

## The platform hierarchy

Bevaya organizes everything inside a three-level hierarchy.

**Organization** — the top-level customer boundary in Bevaya. Your organization contains all of your workspaces, users, and configuration, and is the boundary for access and tenancy.

**Workspace** — a grouping of related AI Agents within an organization. Workspaces let you separate work by team, line of business, or stage of work.

**AI Agent** — the working area that contains flows and their related configuration. AI Agents are where automations are built, tested, and run.

For guidance on choosing the right workspace and AI Agent structure, see [Platform hierarchy](../onboarding/platform-hierarchy.md).

## The build model

You create automations by configuring an AI Agent and defining how it works as a flow of nodes.

**AI Agent** — an automation worker that follows a defined insurance business process, such as submission intake or claim indexing. You can start an AI Agent from a use-case template or build one from scratch.

**Flow** — the node-based definition that determines how an AI Agent processes work. A flow begins with a trigger and continues through the nodes you connect to it.

**Trigger** — the node that starts a flow. It ingests the initial input and injects that data into the flow so downstream nodes can act on it. For available triggers and how to set them up, see [Configure triggers](../build-ai-agents/triggers.md).

**Node** — a configurable step that performs work inside a flow. Nodes are the primary build primitive: outcomes like classification, extraction, matching, routing, human review, and write-back are created by composing nodes. Each node has inputs, configuration, and outputs that downstream nodes consume. See the [Node catalog](../build-ai-agents/node-catalog.md).

**Environment** — the selected runtime context used when testing, publishing, or running a draft. You choose an environment from the environment selector in the builder. See [Configure environments and variables](../build-ai-agents/environments-variables.md).

**Draft and Publish** — the *draft* is the editable state of a flow. *Publish* is the action that publishes the current draft so it can run in production. You can run a draft to test it before publishing. See [Drafts and publishing](../build-ai-agents/drafts-publishing.md).

For the end-to-end build process, see [Build AI Agents](../build-ai-agents/overview.md).

## The runtime model

Once a flow is published and activated, work moves through it at runtime. Publishing alone doesn't start processing — you activate the flow to put its trigger live (see [Drafts and publishing](../build-ai-agents/drafts-publishing.md)).

**Workflow run** — a single execution of an AI Agent against one case: a document package, claim, submission, audit sample, or business event. A configured trigger starts a workflow run in production. You can browse and inspect runs in [Run history](../monitor-review/run-history.md).

**Work item** — a prepared exception task created when Bevaya needs human judgment before completing a workflow action. A work item arrives with context already assembled — the source document, proposed classification, extracted fields, and a recommended match or action — so the reviewer validates or corrects rather than starting from scratch.

**Human review** — the runtime step where a work item is routed to a reviewer for inspection or correction before the flow continues. A Review node in the flow sends an item to human review. See [Human review](../monitor-review/human-review.md).

You can track the status of work across your AI Agent — Failed, In Progress, Review, and Complete — in [Item status reporting](../monitor-review/item-status-reporting.md).

## How the concepts fit together

An **organization** contains **workspaces**, which contain **AI Agents**. You build an **AI Agent** by defining its **flow** — a **trigger** plus the **nodes** that process each case. You test the **draft** in a selected **environment**, then **publish** and **activate** it. In production, the trigger starts a **workflow run**; if a case needs human judgment, the flow creates a **work item** and routes it to **human review**, after which the run completes.

## Where to go next

- [Platform hierarchy](../onboarding/platform-hierarchy.md) — organize your organization, workspaces, and AI Agents.
- [Build AI Agents](../build-ai-agents/overview.md) — create, configure, test, and publish an AI Agent.
- [Monitor & Review](../monitor-review/overview.md) — see how runs, work items, and review behave at runtime.
- [Glossary](../reference/glossary.md) — full definitions for all platform, insurance, and API terms.