Zapier lets AI agents build and deploy Zaps through MCP
Next Gen Zaps moves workflow creation beyond Zapier's visual editor, letting supported MCP clients assemble and deploy automations from plain-language instructions.
What changed
Zapier has documented Next Gen Zaps, an early-access workflow system that lets an MCP client such as a chatbot or coding agent create and deploy a Zap from a natural-language request. The client can select app connections, configure workflow steps and deploy the automation to Zapier.
Zapier says the capability is available by request across Free, Professional, Team and Enterprise plans, with Team access requests requiring an owner or admin.
Why it matters
This changes where automation design can happen. Instead of opening a dedicated workflow builder first, a user can describe the outcome inside an AI environment that supports MCP and hand the resulting workflow to Zapier for execution. That can reduce setup friction for straightforward automations and make Zapier more useful as an execution layer behind AI assistants.
Buyer takeaway
The feature is still early access, so buyers should not treat it as a replacement for Zapier's established visual workflow tooling yet. The practical test is whether agent-built workflows remain understandable, editable and governable once deployed. For teams already standardizing on MCP-compatible assistants, however, Next Gen Zaps is a meaningful expansion of how Zapier can fit into an AI automation stack.