I've seen an issue while testing Agentic where the conversation fails whenever a tool call errors out.
For example, I created a tool that retrieves data from data tables using Genesys APIs. By default, the API throws an error when the requested data is not available which can be pretty much handled in architect flow.

But it fails in Agentic. I've tried multiple variations in the output instructions, but in every case, the conversation ends with an error.

The conversation works as expected when the key field exists in the data table.
Please share your experience around this scenario.
#API/Integrations#ConversationalAI(Bots,VirtualAgent,etc.)------------------------------
Jyoti Sharma
Senior Design Consultant
------------------------------