Zapier expands Next Gen Zaps with branching, recovery and agentic management

Launch Announcement

Today at ZapConnect 2026, we unveiled Next Gen Zaps, a new way to build and run more powerful automations on Zapier, from almost anywhere. This announcement signals a strategic pivot toward integrating AI into the core of Zapier’s automation engine, positioning the platform to handle more complex, data‑intensive workflows.

AI‑First Build Experience

With Next Gen Zaps, you can describe what you want inside MCP‑enabled AI tools like Claude, ChatGPT, or Cursor, and deploy it to Zapier. Or build directly in Zapier using natural language, code, or a visual canvas. The ability to author automations in familiar AI interfaces lowers the barrier for non‑technical users while still offering the depth required by power users.

Enhanced Workflow Capabilities

Branches that split and rejoin to route data down different paths, then merge and continue. Nest loops inside loops to process data and roll it up into a summary at the end. Error handling that keeps going so a failure doesn't stop the whole run. These features directly address the 500‑item loop cap and other limitations that frustrated users of classic Zaps, enabling more robust data processing pipelines.

Resilience and Human‑In‑The‑Loop

Pause mid‑run for a human reply or scheduled follow‑up, then resume automatically. Self‑healing workflows where a monitoring agent watches your runs, diagnoses failures, and can apply or suggest a fix. By embedding pause points and automated recovery, Next Gen Zaps aim to reduce downtime and manual intervention, a significant improvement over the brittle nature of earlier automations.

Access and Availability

Request early access at zapier.com/workflow/early-access . Available on all paid plans (Pro, Team, Enterprise). If you’re on a Team plan, the request must come from an account owner or admin. This rollout strategy keeps the feature behind a paid tier, which may limit experimentation for free‑tier users but aligns with Zapier’s broader monetization model.

ToolScout Analysis - Tradeoffs

While the expanded branching, looping, and self‑healing capabilities promise greater flexibility, they also introduce complexity. Users will need to invest time learning the new visual canvas and AI‑driven syntax, and the monitoring agent’s suggestions may require manual validation. Additionally, the reliance on external AI models raises questions about data privacy and latency that are not yet fully addressed. For teams that already rely heavily on Zapier, the incremental learning curve may be offset by the productivity gains, but smaller teams might find the feature set overwhelming without dedicated support.

Community Engagement

We'd love to hear what you build with Next Gen Zaps. Let us know in the comments! Zapier’s call for community feedback underscores its intent to iterate quickly based on real‑world use cases, a strategy that could accelerate adoption if the platform delivers on its promises.