Most organizations are stuck in “pilot purgatory” despite enthusiasm for artificial intelligence at the leadership level, according to new research from the Asana Work Innovation Lab.
The comprehensive study, which surveyed more than 3,100 knowledge workers in the U.S. and U.K., found that 67% of organizations struggle to roll out AI beyond initial experiments and prototypes to achieve enterprise-wide transformation.
“AI is everywhere in organizations: headlining strategy decks, dominating boardroom discussions, and splashed across marketing campaigns,” the report states. “But despite the hype, most organizations are stuck in pilot purgatory—applying AI piecemeal rather than fundamentally rethinking how work gets done.”
Leadership enthusiasm doesn’t translate to employee adoption
While executives appear bullish on AI implementation, with ambitious roadmaps and proof-of-concept celebrations, the study revealed a significant disconnect with employees who often lack proper guidance and support.
Nearly half (49%) of individual contributors reported their company has done nothing to support them in using AI effectively. This leadership-employee divide manifests in adoption rates, with senior leaders 66% more likely to be early AI adopters than their employees, and managers 38% more likely to use AI weekly than individual contributors.
“The real challenge isn’t getting leadership buy-in—that’s the easier part,” the report notes. “It’s embedding AI into day-to-day workflows so deeply that reverting to the old way of working doesn’t just feel inefficient—it feels unthinkable.”
Five critical chasms to cross
Analyzing AI adoption across more than 112,000 employees and 350 organizations, the researchers identified five key barriers—or “chasms”—that organizations must cross to move beyond AI experimentation:
From AI as a hobby to AI as a habit: Daily AI users were 89% more likely to report productivity gains compared to just 39% for monthly users, demonstrating that value comes from consistent use.
From top-down buy-in to all-in buy-in: Organizations need to bridge skepticism gaps, with individual contributors 39% more likely to be skeptical about AI benefits and 32% more worried about job replacement.
From AI in isolation to AI in context: AI must be embedded into real workflows, not exist as add-ons. Workers were 40% more likely to accept shorter, concise AI summaries than verbose ones.
From AI as a solo act to AI as a team sport: While 49% of AI workflows are built for individual use, they drive only 6% of organization-wide adoption. Cross-functional adoption is key, with users 46% more likely to adopt AI when cross-functional partners use it.
From acquiring users to harnessing influencers: The research found that organizations leveraging internal influencers for AI adoption reached 60-70% of employees quickly, compared to just 30-35% with traditional top-down rollouts.
IT and HR partnership crucial for success
One surprising finding was the importance of collaboration between IT and human resources departments. The report found that IT and HR are 49 times more likely to collaborate on AI workflows than on other business initiatives.
“AI isn’t a technology shift—it’s a work shift. And that demands a deeply human approach,” the researchers noted. “IT ensures AI is scalable, secure, and compliant. HR ensures AI isn’t just implemented—but adopted, understood, and trusted.”
The study was conducted by the Asana Work Innovation Lab in collaboration with researchers from Stanford University and the University of Pennsylvania, analyzing data from more than 112,000 employees across 350 companies adopting AI.
