Organizations are investing heavily in AI training. According to a recent report, 51.6% of organizations are training employees on how to safely use AI agents. The same report also found that only 1% of leaders consider their organizations mature in AI adoption, highlighting the gap between investment and outcomes.
In a recent #shifthappens episode, Dr. Samson Tan, Chief AI Officer at the Institute for Adult Learning (IAL) Singapore, argues that many organizations confuse attendance with adoption and access with application. Employees attend training and gain access to AI tools, but those activities do not automatically change how work gets done.
The shift he advocates is simple but significant: move AI learning out of isolated training events and into the workflows that employees use every day. According to Samson, learning design is what turns AI from a tool people try into a capability they rely on.
Why Training Alone Doesn’t Change Behavior
Training and tool access are important starting points, but neither guarantees that employees will apply AI in their daily work. The assumption is straightforward: If employees know how technology works and have access to it, they will naturally incorporate it into their work.
However, workplace behavior rarely changes that way. Samson states that completing a workshop does not guarantee behavior change, and providing access does not guarantee application. The problem becomes even more pronounced when learning happens separately from the work employees are responsible for delivering. People leave training sessions motivated, only to return to deadlines, meetings, and existing processes that encourage them to fall back into their old, familiar habits. The new knowledge fades because it never becomes part of the workflow.
As he explains, learning often evaporates the moment employees encounter a real-world deadline. If employees cannot see how AI helps them do their jobs better, training becomes another obligation competing for their attention.
How Learning Design Turns AI Use into AI Adoption
For Samson, the problem is not a lack of AI training. It is that organizations often separate learning from the work employees are expected to do. At IAL, learning design starts with real workplace challenges rather than AI tools themselves. Employees bring actual tasks into the learning process and use AI to work through them, making the learning immediately relevant and easier to apply. By embedding learning into the workflow, organizations move beyond teaching AI concepts and help employees develop the confidence and experience needed to use them in practice.
Throughout the discussion, he outlines the principles that help organizations move beyond AI exposure and toward lasting adoption. Together, they show how learning design connects training to application, making AI part of the way work gets done rather than another initiative employees are expected to remember. As Samsonhe describes it, formal training may provide the spark, but learning in the flow of work is what sustains adoption over time.
Stop Confusing Attendance with Adoption
Embed AI into real workflows, so employees apply what they learn, not just complete training and gain tool access. Adoption happens when new skills become part of everyday work, not when a course concludes.
Design Learning Around the Work, Not Around the Tool
Start with the problems employees already have, then introduce AI as part of the solution. Learning becomes more effective when employees can immediately connect AI capabilities to business outcomes.
Replace Output Metrics with Judgment Metrics
Measure how employees evaluate, refine, and apply AI-generated content instead of tracking prompts, logins, or usage frequency. Real capability is reflected in stronger reasoning, decision-making, and execution.
Protect Professional Judgment as AI Scales
Keep accountability with the people doing the work. AI can accelerate tasks and provide recommendations, but expertise, context, and responsibility remain human strengths.
Make AI Part of the Workday
Move beyond one-time learning events. Adoption becomes sustainable when AI is embedded into everyday workflows rather than treated as a separate learning activity.
Turning AI Adoption into Measurable Change
One of the most compelling examples Samson shared came from IAL’s own experience. As AI tools became more capable, curriculum designers faced the same anxiety that many knowledge workers were experiencing. Some questioned whether AI would eventually replace key aspects of their work. Rather than avoiding the technology, IAL created opportunities to explore it within the flow of work through an initiative known as “Tech Thursday,” where teams experimented with AI against real curriculum-development challenges.
The goal was not to learn AI for AI’s sake. It was to improve the work. Teams used AI to accelerate foundational lesson planning, reduce administrative effort, and support curriculum development workflows, while educators continued to apply the professional judgment and contextual expertise that AI could not provide. Rather than replacing curriculum designers, AI handled parts of the process, enabling them to spend more time on higher-order thinking and the human aspects of learning design. When AI was framed as another skill employees needed to learn, it felt like additional work. When it was framed as a tool that could help reclaim time and improve outcomes, adoption accelerated.
That experience also shaped how Samson thinks about measurement. Attendance figures, login statistics, and prompt counts may demonstrate usage, but they reveal little about business impact. Instead, he argues that leaders should focus on reasoning, reflection, iteration, and assisted performance. If organizations only measure activity, they will optimize for activity. If they measure outcomes, capability, and reduced friction, they are far more likely to drive meaningful transformation.
Build on Human Capability
During a visit to Tarragona, Spain, Samson was struck by how the city had evolved over centuries without abandoning its foundations. Modern buildings, businesses, and daily life coexist with Roman structures built more than 2,000 years ago. Rather than replacing what came before, each generation built on top of those prior structures.
For him, that is how organizations should think about AI adoption. The goal is not to discard the expertise, judgment, and institutional knowledge people have developed over years of experience. The goal is to build on those foundations and use AI to amplify human capability. When organizations do that, AI becomes more than a tool that employees learn. Instead, it becomes a natural extension of how their expertise creates value.
Soundtrack of Shift
Samson Tan's Soundtrack of Shift is “A Change Is Gonna Come” by Sam Cooke. The song captures the uncertainty many employees feel as AI reshapes familiar ways of working, but it also reflects the hope and resilience required to navigate that change. For him, learning design helps turn that uncertainty into confidence by embedding AI into daily work and providing the support people need to adapt.
Explore more soundtracks shaping how leaders approach change and transformation today.
Episode Resources
#shifthappens Research: 2026 State of AI
#shifthappens Insights:
- How to Train Your Workforce to Work Alongside AI Agents
- The Illusion of Progress: Why “Doing AI” Rarely Means Deploying It
- AI Literacy Is the New Leadership Imperative for Organizations
#shifthappens Podcasts:
- Build AI That People Welcome
- Turning Hype Into Habit: Making AI Adoption Stick
- AI in Action: From Change to Competitive Advantage
Dr. Samson Tan on LinkedIn
Dux Raymond Sy on LinkedIn
Institute for Adult Learning website
