Unexpected Multiple Agent Execution in Multi-Agent Orchestration in AI Agent Studio
Hello,
I am using AI Agent Studio in Oracle Fusion Sales Cloud to implement a multi-agent orchestration. Each agent has a specific role, and I have added strict instructions in the prompts to ensure that only the relevant agent should be invoked for a given question.
However, we are observing that when a question is asked, other agents are also automatically executing, which:
- Increases response time
- Creates unnecessary calls
- Though the final response is correct, the orchestration is inefficient
Could Oracle clarify:
- Is this expected behavior in multi-agent orchestration?
- Are there recommended settings or best practices to strictly restrict execution to only the intended agent
Tagged:
0