When adopting AI in their operations, executives most often measure AI's return on investment (ROI) in terms of productivity metrics. Yet, employees often can experience that same AI rollout as a threat to their livelihood. This gap in perception explains, in part, why so many AI initiatives stall despite an organization's significant investment.

Indeed, AI transformation succeeds or fails because of human behavior, according to Dr. Gleb Tsipursky, behavioral scientist, CEO of Disaster Avoidance Experts, and author of the forthcoming The Psychology of AI Adoption at Work: From Resistance to Results. And new research from Thomson Reuters backs this up as its 2026 Future of Professionals Report found that among firms with a named AI strategy, 66% of professionals say AI is meeting or exceeding expectations for creating value, compared with just 22% at organizations with no active strategy.

Here are three of Dr. Tsipursky's most insightful recommendations for organizations looking to close the gap between AI strategy and AI results.

1. Trust Is the real bottleneck

Dr. Tsipursky identifies uncertainty as the biggest psychological barrier to AI adoption. "AI adoption moves at the speed of trust," he explains, adding that often executives hear the word productivity, while employees hear a different question in terms of “What happens to my job?” This uncertainty creates resistance, quiet noncompliance, and superficial use of new tools.

To solve for this, organizations need to give employees answers to four specific questions before they will engage with AI in good faith, Dr. Tsipursky says. Employees need to know which tasks will change, which decisions remain human-based, how mistakes will be handled, and what the organization plans to do with the time that AI saves employees. Without clarity on these points, employees assume the worst and act accordingly.

Not surprisingly, the scale of this problem shows up in the data. The Future of Professionals Report found that among professionals whose firm or department has a named AI strategy, 35% of them say that strategy is not visible in their day-to-day experience, and another 17% say their organization has no strategic direction on AI at all.

Dr. Tsipursky offers a success story from his book that illustrates what happens when organizations build trust instead of demanding compliance. He relates how one midsize manufacturer moving forward on AI adoption formed a cross-functional team that included engineers, production managers, quality staff, IT, and HR.

Employees helped design the AI tools rather than having them imposed from above, and the results were measurable. An AI scheduling tool reduced downtime by 14%, and a quality-control system reduced defects by 10%. Also, an employee-built AI assistant cut documentation work by 22%.

These numbers tell only part of the story, however; because the manufacturer's deeper success was psychological, Dr. Tsipursky explains, because employee ownership turned AI from a threat into a tool they wanted to improve. "People support what they help create."

Indeed, this single case captures a principle that applies across industries —participation builds trust, and trust builds adoption.

2. Redesign incentives and workflows

Dr. Tsipursky argues that employees respond to the incentives, workflows, and manager behavior they experience. This means organizations must go beyond telling people to simply use AI more.

Organizational leaders should:

Provide guidance on what to do with the time that AI use saves employees — Protected time stands out as the most valuable incentive. Leaders often ask employees to experiment with new tools while leaving every deadline, meeting, and performance target untouched. Dr.Tsipursky points out that this approach turns AI adoption into unpaid overtime.

Incentivize manager behavior because it carries even more weight than formal incentives — Employees watch closely whether their manager rewards experimentation or punishes the first mistake, and this observation shapes their willingness to engage.

Set up a central team of experts to assist in mapping workflows — Dr. Tsipursky recommends deploying a small central enablement group that provides approved tools and guardrails, and pairing that group with team-level experts who understand the real work on the ground in order to map it. This means identifying where work begins, where time gets lost, what AI can assist with, what humans must decide, how outputs will be checked, and when a case must be escalated.

For example, customer service offers exactly this kind of role-specific redesign. AI can retrieve policies, summarize customer history, draft a response, and document the interaction. The employee remains responsible for checking accuracy, reading emotional context, handling exceptions, and approving the final response. High-risk decisions involving refunds, legal exposure, safety, or vulnerable customers stay with humans.

The talent risk for avoiding effectively addressing this reality could impact AI’s ROI negatively. Dr. Tsipursky warns that without a decision about what happens to saved time, AI becomes a mechanism for increasing quotas —and, not surprisingly, employees learn to hide the efficiencies they discover.

3. Measure what matters and then pace change deliberately

Many organizations track adoption rates and time-saved, but Dr. Tsipursky cautions that these numbers mean little on their own. "Never measure time-saved without asking how that time was used," he explains. "And never measure adoption without asking whether anything important improved."

The data in the Future of Professionals Report underscores this and advises leaders to measure behavior change instead of adoption rates. Leaders also should ask whether professionals can point to something they do differently now because of the AI strategy. The report also describes a predictable pattern in AI rollouts in which early enthusiasm gives way to a period of disappointment and reversion to old habits before real capability takes hold. Reinforcement during this difficult time goes a long way to determining whether organizations push through that middle stretch or mistake it for failure.

In fact, change fatigue presents a separate challenge that many organizations are only beginning to acknowledge. Dr. Tsipursky's guidance recommends that organizations should stay the course strategically while slowing the volume of simultaneous change. Many companies keep launching pilots and changing tools, which creates motion without progress. His recommendation is to choose a small number of valuable workflows, stabilize the tools and guardrails, give teams time to learn and improve, and stop weak initiatives instead of piling new ones on top. The right formula combines consistent direction with disciplined pacing.

As Dr. Tsipursky explains, trust opens the door to adoption, and redesigned incentives and workflows will keep employees walking through it as disciplined measurement and pacing can protect against burnout.

Those organizations that treat AI transformation as a change management challenge, and not a technology rollout, stand a far better chance of capturing the returns they were promised.

You can download your copy of the 2026 Future of Professionals Report here‍