Amazon paper: single agents beat multi-agent setups on enterprise SOPs at 1.4–4.7x lower cost
rohanpaul_ai · x · 2026-11-11
An Amazon research paper holds model, tasks and tools fixed while comparing four agent architectures—single function-calling agent, skill-augmented agent, centralized orchestrator-workers, and a triage swarm—across 13 enterprise SOPs.
Key findings:
- No universal best architecture; the choice is SOP-dependent. Success rates range from under 10% to near 100% per SOP; no architecture is statistically separable in aggregate, and extra structure only plausibly pays off on complex, branching workflows.
- Coordinated architectures must justify their cost: multi-agent and skill-augmented setups spend 1.4–4.7x the tokens and 2.6–5.8x the latency of a single agent.
- For teams automating long, branching business procedures, a single agent is usually the best default; the paper also introduces a skill-definition quality rubric for building matched skill corpora.
Related event: Amazon Paper: Multi-Agent Architectures Cost More, Gain Little(2 posts)→
More from coding & agent
- FluentCRM MCP Server lets you manage marketing automation from Cursor or Claude — modelcontextprotocol · 2026-11-11
- Elvis Saravia: agents will subsume traditional software — and can even cut code bloat — omarsar0 · 2026-11-11
- Using a third-party inference provider with your coding harness hands it shell access to your machine — TheZachMueller · 2026-11-11
- Splash One: flight combat sandbox built with Opus 5.5 in PlayCanvas — willeastcott · 2026-11-11
- Next.js 16.5 delivers 3.3x faster builds for a leading AI lab — cramforce · 2026-11-11
- Plurai launches free agent simulation testing inside Claude Code and Cursor — PrajwalTomar_ · 2026-11-11