The most expensive mistake in multi-agent design is reaching for it by default. Anthropic's "Building effective agents" opens with exactly this warning: the most successful production implementations use simple, composable patterns, and complexity should be added only when it demonstrably pays. A single well-run loop is cheaper, faster, and radically easier to debug — one transcript, one context, one place to look when it breaks.
When multiple agents actually win
Three conditions justify the jump:
- Genuine independence. Subtasks that don't need to talk mid-flight — parallel research angles, per-file migrations, independent review lenses. Anthropic's multi-agent research system is the canonical worked example: a lead agent decomposes a query, parallel subagents each search with their own context window, and only distilled findings flow back. Their honest accounting: it outperformed single-agent on breadth-heavy research and consumed far more total tokens — the win is capability and wall-clock, never cost.
- Conflicting contexts or tools. A safety reviewer shouldn't share a context with the generator it reviews; a sandboxed code-runner shouldn't hold production credentials. Separation here is a control boundary, not an optimization.
- Specialist prompting. When one system prompt would have to be three things at once, three scoped agents each stay sharp — the pattern frameworks like AG2 (group chats, nested conversations, handoffs) and the OpenAI Agents SDK (agents-as-tools, handoffs) both make first-class.
The orchestration patterns that recur
- Supervisor / worker: one coordinator owns decomposition, contracts, and synthesis. Workers never see each other; all integration risk concentrates in the merge step — which is where to spend your testing budget.
- Pipeline: each agent transforms and passes on. Cheap to reason about; fragile when a middle stage silently degrades, so validate at stage boundaries.
- Debate / panel: independent attempts judged against each other. Buys robustness on wide solution spaces; costs a multiple of every token.
Whatever the topology, the handoff is the product: a contract carrying goal, evidence, current state, ownership, and completion criteria. Free-prose handoffs are where multi-agent systems lose the detail that mattered.
Budget the coordination tax
Coordination messages, duplicated context, and synthesis passes are pure overhead — cap fan-out, cap recursion depth (an agent spawning agents that spawn agents is an outage generator), and give every worker an explicit token/time budget. If the overhead exceeds the parallel win, the correct architecture was one agent all along.
Sources: Anthropic — Building effective agents · Anthropic — Multi-agent research system · AG2 documentation · OpenAI — Agents guide.
Related: subagents, planning & decomposition, agent cost control, context rot.