Share with your CIO
A new Asana survey of more than 4,500 CIOs, AI leaders, and workers across the U.S. and U.K. reveals a sharp split in agentic AI adoption rates inside enterprises. 68% of executives use three or more AI agents weekly, versus just 36% of workers. Fewer than one in four workers say available agents actually address their highest-priority tasks. And 38% of workers report that more than a quarter of all agent work is duplicated or overlapping, a coordination tax that compounds as agent counts rise.
What this means for your business
The gap here isn’t enthusiasm. It’s exposure. Executives operating at portfolio altitude see agents as a net positive because they’re mostly insulated from the daily friction: the tool-switching, the conflicting outputs, the ambiguous accountability that workers absorb directly. If your organization is in active agent rollout, the workers who seem “resistant” may actually be your most accurate signal. Their hesitation is calibrated to real dysfunction, not fear of change.
Asana’s CPO frames the fix as shared context and clearer ownership, which is directionally right, even as Asana sells exactly the workflow orchestration layer that would sit between agents and workers. The more durable insight is that agent sprawl, many agents deployed without a coherent operating model, creates a coordination overhead that erodes the productivity gains the investment was supposed to generate. One in four agents already running unmonitored inside enterprise environments, per a separate Workato report cited here, means that sprawl isn’t hypothetical. It’s already the default.
Shift the measurement frame before the next budget cycle. Counting agents deployed, or even agent interactions logged, tells you almost nothing about whether work is finishing faster or better. The leading indicator that actually matters is whether workers can act decisively on agent outputs without a reconciliation loop. If that loop exists and nobody owns it, the agents are producing costs, not value. That’s the renewal worth scrutinizing: not the agent platform, but every downstream workflow that was quietly redesigned around it.
Based on reporting from Agentic AI gap persists between executives, workers, originally published 2026-10-16 16:39:00.
