Anthropic: one AI agent beats sub-agent fleets, about twice as fast

Anthropic's recommendation comes from hands-on experience across multiple enterprise deployments, where its Applied AI team found that a single agent using skills frequently beat both monolithic prompt designs and sub-agent setups on quality. One company had a dozen specialized sub-agents inside its customer-facing agent, which made a consistent brand experience hard to keep and hurt usability with high latency.

The company describes a familiar drift it calls "Conway's Law" for agents: the architecture becomes a copy of the org chart, with each team shipping its own sub-agent until hundreds of overlapping agents pile up. What follows is a "Frankenprompt," where every team adds rules to the system prompt that end up contradicting each other. Anthropic says a clean system prompt, four or five lean data tools, and fixed output guardrails cover much of the respective use case, and it cut the Claude Code system prompt for newer models by over 80 percent, partly because skills and tools now load only when needed.

Anthropic's findings are consistent with recent research. A study by Vals found that agent teams can be faster than single agents but don't deliver higher quality, and scaling from ten to 30 or 100 agents brought no quality gains but significantly higher costs. OpenAI developer Eric Provencher has also questioned the value of agent swarms, calling them "almost certainly" wasted money.