A multi-agent AI workflow takes about 2 minutes for simple answers

A using takes about 2 minutes to answer even a simple question. The flow runs through input guardrails, a planner agent, several sub-agents, a response builder, and .

NeMo is used for the input and . The system works, but every layer runs before the final response, so the full wait time becomes long.

The main production questions are how to reduce end-to-end response time, whether guardrails should run one after another or in parallel, and how to stream answers while agents are still working. The example is shared as a Smart Bill Book workflow agent diagram on GitHub.

Key points

  • The workflow uses in a .
  • The steps include input guardrails, a planner agent, sub-agents, a response builder, and .
  • A simple query can take about 2 minutes because all layers run before the final answer.
  • s must decide whether guardrails should run in sequence or in parallel.
  • Streaming can reduce the felt wait time while agents continue working.
Read original