Skip to main content
Workflows in Forest let operators run multi-step operational processes, KYC reviews, refunds, customer onboarding, supplier onboarding, incident response, directly from the operations UI, without scripts or developer involvement. Each workflow combines data, actions, AI-powered steps, and decisions into a single guided sequence with full audit trail. For developers building or configuring workflows, see Workflows (process). This page is for operators executing them.

What a workflow does

A workflow chains multiple operations into a single automated sequence. Examples:
  • KYC review: load customer record → fetch KYB documents → AI summarizes red flags → operator decision → trigger compliance action → log result
  • Customer refund: validate eligibility → check balance → trigger Stripe refund → notify customer → close ticket
  • Supplier onboarding: create account → request documents → AI validates submission → escalate exceptions → finalize contract
Each step is recorded, every decision, every input, every action triggered, producing a full decision trace operators (and auditors) can replay.

Where workflows are triggered

Workflows can be executed from:
  • The List View (on a single record)
  • The Summary or Details view of a record
  • A Workspace
The workflow opens in a guided panel. The user works through each step; the workflow context accumulates as they go.

AI-powered steps

The Get Data, Update Data, Trigger Action, and MCP Task steps surface the relevant data or actions automatically based on the workflow’s context. Each completed step adds to that context, helping the AI suggest better defaults and pre-fills for subsequent steps, for example, pre-filling fields when updating a record, or composing an email with the right context when notifying a customer through a tool like Zendesk. Every AI-powered step provides a Handle Manually option, letting users bypass the AI suggestion if it isn’t relevant or if the AI hits an error. Subsequent steps then lack that context, which may affect downstream steps.

Automated steps

Some steps run without user input, typically AI-powered steps and certain Decisions. If the AI lacks context to make a confident decision, or if a third-party system is unresponsive, the user is prompted to take over (same as Handle Manually).

Saving and resuming

A workflow can be paused at any time. The user is prompted to Save or Abort it. Saved workflows can be resumed by another user, after any length of time:
  • They resume automatically when triggered from a record’s Summary View or a Workspace
  • They appear under To Continue when the workflow is accessed from the List View or Record View
This is how queue handoffs work, one operator starts a case, leaves it for review, another picks it up.

Revising steps

Workflows reduce mistakes in complex processes, but errors still happen. To correct something, go back to a completed step and click Revise to redo it. Since workflows can branch based on user choices, revising a step cancels any steps completed after it, ensuring the workflow follows the correct path based on the updated input.

Audit trail

Every workflow execution is recorded. The full sequence of steps, decisions, inputs, and outputs is replayable in read-only mode from the workflow list on the record. Every interaction is also logged in your Activity Logs, accessible from the Reports tab or via the public API for audit and compliance reporting.

Permissions

Workflow execution is governed by the standard role and team permission system. A user can only:
  • See workflows their team has access to
  • Trigger steps for which their role has the required permissions
  • Access data and actions inside steps based on their scopes
See Roles & permissions for configuration.