Glean's founder attributes rising enterprise AI bills to complex multi-step task chains, causing token costs to double every 45 days against a 5% productivity lift. The company's routing and context architecture achieved 30% fewer tokens and 2.5 times more preferred answers than other Multi-step Chain Processing (MCP) tools. This efficiency gap widened for harder tasks, with Glean's reported context-layer win rate increasing from 66% to 73%.
Enterprise AI agent deployments are driving token consumption faster than price declines, shifting cost optimization to workflow architecture and context management.