Balyasny's AI Agents Cut Deal Analysis From Days to Hours

Balyasny Asset Management, managing $38 billion in assets, has shifted from AI-assisted search to AI-driven work in 2026. Merger-arbitrage analysis that once took three to five days now takes under one, with agents completing multi-step workflows in roughly 30 minutes. The firm tested Claude Fable 5 on thousands of real financial tasks, achieving 89.4% accuracy versus 86.1% for its prior model. BAM built its own internal platform, BAMAgent, to govern agentic deployments with strict data boundaries, least-privilege access, and human review for material outputs. The firm treats safety as an ongoing operating question, not a one-time model selection. A more capable model does not automatically receive broader system authority.