How to configure context and escalations for agents - Course reflection

Learning is better when it’s shared—and you’ve just built your first agent with escalations!

With this occasion, we invite you to share your “A-ha!” moments with us and fellow learners. If it’s easier for you to take some time and reflect on this lesson with some guidance, don’t hesitate to refer to one or more of the following questions:

  • What did I learn and did not expect?
  • What was difficult for me and how did I overcome it?
  • What is the most useful thing that I learned in this lesson?

Feel free to share with us and your peers whatever comes to your mind!

With this lesson, I had a few “A-ha!” moments that I’d like to share.

I learned how to configure context and escalation in an agent, which I didn’t expect at first.
It wasn’t difficult for me to understand or apply the concepts.
The most useful thing I learned was how to build a multi-agent solution for handling customer complaints.

Thank you for the great lesson!

My “A-ha!” Moments While Building an Agent with Escalation

This lesson gave me several key takeaways that really clicked for me!

:glowing_star: What I learned and didn’t expect:
I didn’t realize how seamless it could be to combine context grounding with escalation logic. Learning how to handle unclear cases by involving a human supervisor through the Action Center really opened my eyes to how flexible and intelligent these agents can become.

:hammer_and_wrench: What was difficult and how I overcame it:
The process itself was well-structured, so I didn’t find any major difficulties. Understanding how to structure the system and user prompts properly was key—and once I got that, everything flowed.

:bullseye: Most useful thing I learned:
Building a multi-agent system where a classifier agent hands off to a complaint resolution agent—with escalation for unclear cases—was a game-changer. It mirrors real-world enterprise use cases perfectly.

:speech_balloon: To my fellow learners:
If you’re still working through the lessons, feel free to reach out. I’m happy to help if you hit a roadblock. Let’s grow together!

– Roopa

The only “issue” for now has been the differences in the Studio web Enterprise

1 Like

That’s true. In fact, you cannot add multiple agents into the single solution, which made it hard to catch up with the exercises.

On the other hand, if you import a complete solution, you can add agents and apps. Strange.

I was unable to follow along with “Build an escalation app for your agent” section as I did not see the “Generate page” button when trying to create the escalation app.

1 Like

Same issue happened to me, I tried to use the Escalation app but same behavior. Page was not created so I created one page manually but when I tested it didn’t open or redirect to the escalation app.
I could see an error message “Actions is not enabled for this tenant, Actions requires the UiPath automation Cloud”

I’m currently following the Agentic Automation training module specifically the lesson “Build an escalation app for your agent.”

While trying to create the escalation app (after configuring the Action Schema), I’m getting this error message:

“We’re currently facing an issue with the Autopilot service which is preventing us from processing your request. Please try again later, or contact your administrator for further assistance.”

Trace Id: d3b17331cb9d4c649099092f7b21bc68

Even when I try to manually create an app by going to Apps → Create App (without using Autopilot), I still receive an error saying:

“App addition is in progress. Please check after some time.”

It looks like the page creation or Autopilot process fails at the backend.

I’ve already tried:

  • Logging out and logging back in
  • Clearing cache and cookies
  • Trying a different browser (Chrome and Edge, also in Incognito mode)
  • Creating a new solution from scratch

But the issue persists.

Could this be related to any tenant/organization-level configuration, region-specific outage, or permission setting?
For example:

Do I need to enable any specific Autopilot or App service capability in my organization settings?
Is there any dependency on Action Center licensing or folder permissions for this step?
Or is this a known temporary issue in the cloud environment?

Any guidance or workaround would be appreciated