In this workshop, you’ll build a multi-agent customer support system using AI Agent Builder + ChatGPT.

Instead of one generic chatbot, the system will detect the customer’s intent and route the request to a dedicated specialist agent, such as:

• Returns Agent — policy-aware checks, eligibility, step-by-step return guidance
• Plan Change Agent — plan options, constraints, confirmations, and next steps
• Cancellation Agent — compliant language, final confirmation, retention-safe responses, and escalation paths

Guardrails and Rules (Reliability + Safety)

We’ll add guardrails and strict rules so the agent behaves consistently and avoids risky or incorrect outputs, including:

• Policy boundaries (what the agent can/can’t promise)
• Refusal and escalation rules (when to hand off to a human)
• Data handling rules (PII-safe responses and safe wording)
• Response structure rules (consistent formatting and clear outcomes)

Python Implementation (Portfolio-Ready)

Finally, we’ll translate the complete workflow into Python, so you leave with a clean, reusable project you can extend and deploy for real customer support use cases.