Building the Agentic Enterprise

3

Single-Agent and Multi-Agent Systems

This chapter asks a deceptively simple question: how many agents do you actually need? Before adding more roles, handoffs, and orchestration, let's establish when a single agent is enough, when decomposition genuinely earns its complexity, and what that extra coordination costs once the system reaches production.

In Chapter 2, we established the Reason-Act-Observe loop as the engine of agency and introduced the single-agent MarketDataAgent. A single agent is easier to observe, cheaper to operate, and often capable of handling far more than teams initially expect. But some enterprise workflows eventually expose a cognitive ceiling: too many responsibilities, competing objectives, different permission boundaries, ...