DMO Geek
Back to Use Cases
Operations & AutomationAdded February 10, 202667 proposals processed

Ops Proposal & Approval System

Submit work proposals, approve them, and auto-spawn AI agents to execute — all with human oversight.

Overview

A closed-loop system where AI agents propose work, the team lead approves or rejects via dashboard, and approved missions auto-spawn execution agents with full audit trails. Every action requires explicit human approval — no autonomous task execution.

How It Works

Agents submit proposals to Supabase with context, scope, and expected outcome. The team lead reviews in the dashboard and clicks approve or reject. Approved proposals trigger a dispatch system that spawns a mission-specific subagent with the proposal context. The executing agent reports progress via heartbeats and files a final report to Notion.

Tools Used

SupabaseNext.js dashboardOpenClawNode.js

Outcome

Processed 67 proposals in February 2026 with 49 approvals. The system maintains human control while eliminating manual task assignment and progress tracking overhead.