All templates
FeaturedAI#HITL#Agent#Human Review#Support#Approval

HITL Support Reply Agent

Draft a customer-facing support response, pause for human approval, then continue with the reviewed reply.

Workflow at a glance

The full canvas, before you import it

Click any node to see its config.

#HITL#Agent#Human Review#Support#Approval

Click a node to select it — same as the Heym editor; the panel shows its settings.

5 nodes · Free & source-available

HITL Support Reply Agent

This template shows Heym's Human-in-the-Loop (HITL) flow on an Agent node. The agent drafts a customer response, calls request_human_review, exposes a review link on the HITL branch, then resumes after the reviewer accepts, edits, or refuses the draft.

What this workflow does

  1. CustomerRequest provides a sample support escalation
  2. DraftReplyAgent prepares a customer-ready response and requests human review before finalizing it
  3. NotifyReviewSlack sends the pending review URL, summary, and Markdown draft to Slack from the HITL branch
  4. ApprovedReply returns the final reviewed reply and decision metadata after the run resumes

Use cases

  • Customer support replies that need manager approval
  • Refund, escalation, or legal-sensitive response review
  • Portal workflows where a human should approve an agent action before it reaches a customer

Setup

Connect an LLM credential in the Agent node and a Slack credential in NotifyReviewSlack. Run the workflow, open the generated review URL from Slack, then accept, edit, or refuse the draft to resume execution.

Notes

HITL works with text-mode Agent output. Keep JSON output disabled on the Agent node when using this template.

How to import this template

  1. 1Click Import → Copy JSON on this page.
  2. 2Open your Heym and navigate to a workflow canvas.
  3. 3PressCmd+V/Ctrl+V— nodes appear instantly.
  4. 4Add your API keys in the node config panels and click Run.
More workflow templates
View all templates
Heym
incident analysis · production AI
Observed across 100s of AI rollouts

AI workflows don't fail because of prompts.
They fail because of orchestration.

symptom · glue code01
5 tools
Scripts, vector DB, approval bot, tracing, browser runner — none of them talk.
symptom · visibility02
~0%
Observable behavior across the stack. Debugging is guesswork.
with heym · one runtime
1 canvas
Agents, RAG, HITL, MCP, traces & evals. Self-hosted. Observable.
AI-Native RuntimeProduction-Grade
github.com/heymrun/heym