An AI assistant configured with specific instructions, tools, and connections. Agents handle incoming documents, answer questions, and push data to external systems. Each Agent is set up by an admin and can be scoped to specific users or channels.

A file (purchase order, invoice, sales order, etc.) submitted to Fluent for AI extraction and processing. Documents move through a status lifecycle — Draft → Pending → Approved → Processed — before being pushed to your connected system. See Understanding document statuses for details.

A conversation between a user and an Agent. Threads are searchable, shareable, and visible to team members with access. Your recent threads appear in the sidebar; the full list is under Threads.

An integration to an external system — such as Epicor, SugarCRM, Salesforce, or E3 — that Agents use to read and write data. Connections are configured by an admin and shared across the workspace.

Your shared Fluent environment. Agents, documents, connections, and members are all organized at the workspace level. Users can belong to multiple workspaces and switch between them from their account.

A team member designated to review and sign off on documents before they can be processed. If a document requires approval, it sits in Pending status until all required approvers act on it.

An action an Agent can take during a conversation — such as fetching data from your ERP, sending a notification, or running a calculation. Tools are enabled per Agent by an admin.

An event that automatically starts an Agent conversation — for example, an inbound email, a scheduled time, or a webhook. Triggers remove the need for a user to manually initiate the flow.

A collection of reference documents (product specs, pricing sheets, policies, etc.) that an Agent can search during conversations. Admins upload files to a Knowledge Base and attach it to one or more Agents.