Agentic AI Foundations / Lab 1C
← Course Hub
LAB 1C · OPTIONAL · SHOW-AND-TELL · 35 MIN

Orchestrate a Month-End Close Automation

Advanced & optional: combine your Lab 1A + 1B agents into one end-to-end process with Quick Automate — decision branch, document generation, a human approval gate, then commit · deploy · monitor. Enterprise-only; watch or follow along.

🎼 What This Lab Demonstrates

This lab combines the Policy Knowledge Agent (Lab 1A) and the Break Triage Agent (Lab 1B) into a single orchestrated month-end close automation — showing what a production-grade pilot looks like once it graduates past a single Chat Agent.

Enterprise-only: Quick Automate's commit · deploy · monitor lifecycle requires an enterprise tier not available in every workshop environment. Follow along with the instructor demo rather than building this hands-on unless your environment supports it.

🪜 The Orchestrated Flow

  1. Trigger: Scheduled run at close-cutoff, or manual trigger by the Controller
  2. Policy Check: Policy Knowledge Agent confirms which close-calendar deadlines apply this cycle
  3. Break Triage: Break Triage Agent classifies all outstanding reconciliation exceptions
  4. Decision Branch: If any high-severity break remains unresolved past cutoff, escalate to the Controller before proceeding
  5. Document Generation: Combined close status pack — outstanding breaks, escalations, and citations
  6. Human Approval Gate: Controller reviews and approves the pack
  7. Commit · Deploy · Monitor: The approved automation is committed to production, deployed for the next cycle, and monitored for drift or failure

🧭 Why This Matters for Your Pilot Canvas

Module 7's capstone asks you to design a real pilot. This orchestration pattern — multiple single-purpose agents, a decision branch, a mandatory human gate, and a monitored production lifecycle — is the template most real finance pilots should follow rather than one large, do-everything agent.