Configure a beta agent with a bounded goal, named tools, and explicit execution limits.
Workflow & automation · Beta / preview · Documentation checked 2026-10-04
Make scenarios and Make AI Agents (New); connected model providers have their own terms.
Scenario limits vary by plan. AI Agents (New) is in open beta; custom AI providers require a paid plan.
The worked example is illustrative and has not been tested in a signed-in product account.
What it does and how it connects
Configure a beta agent with a bounded goal, named tools, and explicit execution limits.
Make AI Agents (New) is in open beta. It can use module tools, scenarios, and MCP, with configuration for execution limits and output. A goal directs the agent; the tool definitions and connections determine what it can actually do.
Walkthrough
Use the product named in the access panel. The steps below combine documented controls with the editorial example.
- Add the documented AI Agent (New) module in a sample scenario.
- Choose Make’s provider or an eligible custom provider connection.
- Define a bounded goal and add only named read-only tools with clear input / output contracts.
- Set supported step / time bounds, test, and inspect activity before enabling consequential tools.
Worked example
A sample FAQ agent looks up a workshop session and produces a draft answer. It should not have registration or email tools during the pilot. Tool names should describe specific operations rather than vague general assistance.
Answer workshop questions with the approved FAQ lookup tool. Return the answer, source reference, and unresolved details. Stop on missing or contradictory evidence. Do not invent dates or fees. Read only; if the requested task needs a write, return needs_review instead.
Inspect and verify the result
Inspect selected tools, mapped inputs, output, and execution metadata. Test irrelevant questions and tool failures. Confirm the actual configured maximum steps and timeout rather than relying on prose alone.
Limits and troubleshooting
Beta behavior and access can change. Make’s provider and custom-provider paths have different plan and billing requirements. A prompt saying ask for approval does not create an external enforced approval boundary.
Data, permissions, and actions
Model context goes to the selected provider and tools act through connected accounts. Review MCP authority and script / scenario side effects. Restrict equivalent write tools and place consequential actions behind a separately validated review gate.
Related reading
Knowledge, conversation IDs, and structured output
Distinguish reusable files from run inputs and request an output downstream modules can validate.
Fixed AI triage with a review filter
Classify synthetic requests while keeping authoritative review separate from AI output.
A bounded beta knowledge agent
Use the beta agent to retrieve approved workshop evidence and produce a validated response draft.
Permissions and meaningful approvals
Decide what a system may read, what it may change, and when someone needs to authorize a concrete action.
Sources and verification
Documentation checked 2026-10-04. Examples are illustrative unless described otherwise.
Make AI Agents (New) configuration · Make
AI Agent best practices · Make