Participant's Handbook¶
Thank you for joining the UiPath Advanced Agentic Automation Course!
This handbook is your practical guide for the hands-on exercises. It walks you through each step, explains what's happening and why, and helps you build something real by the end of the session.
Training Environment
- Platform: cloud.uipath.com/tpenlabs
- Tenant: AgenticAdvancedWorkshop
Work inside your workshop VM, and make sure the AgenticAdvancedWorkshop tenant is selected in the top right corner before you start. If you can't reach the environment, raise a hand.
What You'll Build¶
This course has two exercises. Each one starts with its own setup lesson, so you can begin with either.
1. First Notice of Loss Agent¶
An insurer receives a First Notice of Loss (FNOL) claim together with the homeowner's policy it refers to. You'll have Codex build a low-code agent in Studio Web that reads both PDFs, runs five eligibility checks, and returns a decision, an escalation flag and a summary.
Then you'll turn a set of test documents into an evaluation set and let Codex run the loop: run the evaluations, find what fails, fix the prompt, and run again until the scores hold.
Start First Notice of Loss Agent →
2. Support Ticket Triage¶
You'll build a support-ticket triage automation for HomeFix in Maestro Flow, with Codex doing the building. An inline agent classifies each ticket and drafts a reply grounded in the HomeFix knowledge base, and a human reviews the draft in Action Center.
Then you'll do the part that matters: feed your review decisions and trace feedback back to Codex, and let it build a second and third version that route urgent tickets, drop noise, and finally auto-resolve the safe cases. Six prompts, three versions, one scoreboard.
Use the navigation at the top to access the lessons!
Something worth holding onto as you work through the exercises
The workshop is not "build a flow". It's an improvement cycle: run real examples, collect human review decisions, leave trace feedback, revise, and add automation only where the evidence supports it.