AI Agents Cross the Chasm, but Oversight Costs Erode Gains
Event summary
- 86% of B2B software buyers purchased AI agent software or services in the past year, up 25% YoY.
- 31% of employees spend 6+ hours weekly bot-sitting, with 45% of organizations losing >25% of expected productivity gains to oversight.
- 38% of organizations run 100+ agents, yet 47% manage them informally or decentralized.
- 84% of buyers say agent capabilities will influence 2027 software choices, with 34% considering them a primary factor.
The big picture
AI agents are transitioning from pilot programs to production, with 66% of B2B buyers now deploying them at scale. However, the 'Distrust Tax'—where oversight and rework consume up to 25% of expected gains—is slowing mainstream adoption. The shift is reshaping SaaS pricing models and forcing organizations to rethink governance frameworks as agent fleets grow. Healthcare leads spending with 57% of organizations allocating $1M+ annually, while technology and retail see the biggest rise in $5M+ budgets.
What we're watching
- Trust Dynamics
- How the 'Distrust Tax' will evolve as organizations demand more proof of agent reliability in real workflows.
- Governance Scaling
- Whether decentralized agent management can sustain productivity gains as fleets grow beyond 100 agents.
- Pricing Models
- The pace at which pay-as-you-go models displace traditional seat-based pricing for AI agent software.