Only 5% of Companies Are Closing the AI Productivity Gap. Asana Thinks It Knows Why.

Asana has released new research revealing a striking disconnect in enterprise AI adoption: while roughly 75% of knowledge workers now use AI tools at work, only 5% of companies are seeing meaningful productivity improvements as a result.

The Details

The findings point to a structural problem, not a technology problem. Asana argues that most organizations are layering AI tools on top of broken workflows, which means the inefficiencies are still there, just automated.

Key findings from the research include:

• Three out of four knowledge workers use AI regularly on the job
• Only 5% of organizations report closing the gap between AI adoption and actual productivity gains
• The gap is not narrowing over time; according to Asana, it is widening

Asana's position is that AI amplifies whatever system it operates within. If your coordination, task management, and communication processes are fragmented, AI makes fragmented work happen faster. It does not fix the underlying structure.

"The problem isn't AI adoption. The problem is that companies are adopting AI without redesigning the work around it."

The report emphasizes that the companies seeing real gains are those that rearchitected workflows before or alongside deploying AI, rather than bolting tools onto existing processes.

Why This Matters for Service Providers

For MSPs and telecom resellers, this research has direct implications for how you position and deploy AI services for your clients. Selling AI tools is not the same as delivering AI outcomes. If your clients adopt AI voice agents, automated ticketing, or intelligent call routing without adjusting their underlying processes, they will likely join the 95% who see little to no measurable return.

This is both a warning and an opportunity. MSPs who take a workflow-first approach when deploying AI solutions will differentiate themselves from vendors who simply flip on a feature and walk away. If you are helping clients add AI voice agents to their service stack, the implementation conversation needs to include process redesign, not just technical setup.

The providers who will win long-term are those who deliver outcomes, not just access to tools.

Looking Ahead

Watch for enterprise buyers to become more skeptical of AI ROI claims over the next 12 to 18 months, which means service providers need to build measurable outcome frameworks into every AI engagement before the market demands it.

For the full story, read the original article on UC Today.