Let the agent do things mid-conversation, like looking up an order or calling your own API, with secrets, approvals, and an audit trail.
AI Actions let the agent take an action during a chat instead of just answering. It can look up an order, find a contact in your CRM, or call your own API and use the result in its reply. There is no trigger to configure: the model decides when to run an action based on the conversation. You decide what actions exist and what they are allowed to do.
Open Capabilities, then AI Actions.
AI Actions are available on Growth and above, with a monthly cap on action calls. Free and Starter do not include them. See Plans.
Two kinds of action
Marketplace apps
Pre-built actions for services we support, like looking up a Shopify or WooCommerce order or finding a HubSpot contact. You connect the app once for your workspace (see Integrations), then enable the actions you want. The agent handles the rest.
Actions that depend on a connected store show up first when that store is connected, so the relevant ones are easy to find.
Custom HTTP actions
A custom action calls any HTTP API you point it at. You define:
- The parameters the agent should collect or infer (for example an order number)
- The request: method, URL, and headers
- How to turn the response into an answer
You can test an action with sample inputs before you turn it on, so you see exactly what the agent will send and get back.
Secrets
APIs need keys, and you should never paste a key into a prompt. Store them as secrets in workspace settings, then reference them in an action's headers as {{secret.NAME}}. Secrets are encrypted, only ever used in request headers, and are stripped from the action's audit log. See Workspace settings.
Approvals
Some actions should not run without a human saying yes, like issuing a large discount. Mark an action as requiring approval, and when the agent wants to run it, the request lands in your Inbox as a card. A teammate approves or denies it, and the visitor waits for the result.
The audit log
Every action run is recorded: what ran, the inputs, a trimmed excerpt of the response, and how long it took. Secrets are never stored in the log. This gives you a clear history of everything the agent did on your behalf.
Usage
Action runs count against your plan's monthly cap. The AI Actions page shows a usage meter and your top actions so you can see what is getting used.