Orchestration vs a user facing agent

[piffy76]

What are differences between using an agent to govern multiple sub agents and using orchestration, such as ConcurrentOrchestration to execute multiple agents? With the former, I know I can instruct the governing agent. With the latter, I don't see a way to instruct the orchestrator. What is the purpose of using Orchestrator in contrast to Agent? When do I use which?

Below is an Orchestration example where I can't give it any instruction.

ChatCompletionAgent physicist =
            this.CreateAgent(
                instructions: "You are an expert in physics. You answer questions from a physics perspective.",
                description: "An expert in physics");

ChatCompletionAgent chemist =
    this.CreateAgent(
        instructions: "You are an expert in chemistry. You answer questions from a chemistry perspective.",
        description: "An expert in chemistry");

ConcurrentOrchestration orchestration =
            new(physicist, chemist)
            {
                ResponseCallback = monitor.ResponseCallback,
                LoggerFactory = this.LoggerFactory
            };

        // Run the orchestration
        string input = "What is temperature?";
        OrchestrationResult<string[]> result = await orchestration.InvokeAsync(input, runtime);

[moonbox3]

Adding @TaoChenOSU to help (code sample is dotnet, but same concept exists for python).