Most organisations spend between £400–500 billion per year on corporate training globally. Yet 74% of senior leaders believe their companies lack the skills they need to compete. This is remarkable. It suggests that current training approaches simply aren't working.
Why? Because the model is fundamentally broken. The traditional corporate learning model assumes that you can extract people from their work for a day or a week, teach them something in a classroom, and have them apply it when they return. But that's not how learning actually works. Research by BuildEmpire (2025) found that 86% of employees learn by figuring things out on the job. Not in classrooms. Not in e-learning modules. In real work, solving real problems, with support from peers and managers.
Bersin's February 2026 research compared traditional learning approaches to AI-first learning programmes. The finding was striking: AI-first learning teams are outperforming even the most advanced traditional learning teams from 2022. But what makes a learning programme "AI-first" isn't primarily the tools. It's the approach: learning embedded in work, supported by micro-resources, measured by performance outcomes.
Learning in the Flow of Work
What does learning in the flow of work actually look like?
Example 1: Just-in-time resource support. A team member is about to use a new AI tool for the first time. Instead of a two-hour training session weeks earlier, they have access to a three-minute video showing exactly how to use the tool for their specific situation. This is learning at the exact moment of need.
Example 2: Structured on-the-job learning challenges. Instead of a classroom session on "leading through change," a manager is given a structured challenge: "Over the next two weeks, notice one moment when your team is resistant to change. Have a conversation with them about what's driving the resistance. Document what you learn. Share with your peer group." This is learning by doing, with structure and reflection.
Example 3: AI-supported performance feedback and coaching. A manager receives feedback that she needs to develop her listening skills. Instead of a training course, she gets: a daily two-minute reminder of a listening technique to practise, AI-powered analysis of her last few team meetings, and monthly coaching with focus on specific development. The learning is woven into her actual work.
Shifting Your L&D Investment
This requires a fundamental shift in how organisations think about learning:
From event-based to continuous. Instead of an annual training programme for each topic, create a continuous learning experience. Monthly micro-modules, weekly reinforcement, real-time support.
From classroom to embedded. Instead of pulling people out of work to learn, embed learning in the workflows where they work. Resources on hand. Support available. Learning happens whilst work happens.
From completion to capability. Stop measuring by "how many people completed the training." Start measuring by whether capability actually improved and how that translated into performance.
From generic to contextual. Stop teaching generic leadership principles in a classroom. Instead, give people learning resources tailored to their specific context, challenges, and role.
From expert-led to peer-supported. Instead of centralised L&D teams designing all learning, create structures for peer learning. Communities of practice. Peer mentoring.
Try This
Identify the top three ‘moments of need’ in your team's week. When are people most likely to struggle? For each moment, create a micro-resource: a three-minute video, a one-page guide, an AI-powered prompt. This is learning embedded in work.
Replace one classroom session with a structured on-the-job learning challenge. Instead of a workshop on ‘managing remote teams,’ try: ‘Over the next two weeks, have three conversations with direct reports about what helps them do their best work remotely. Document what you learn. Share with peers.’
Shift one metric from completion to capability. Instead of measuring ‘how many people completed the training,’ measure ‘did capability improve?’ Did decision quality increase? Did speed improve? Did engagement rise?
References
Bersin, J. (2026) AI in Corporate Learning: The definitive analysis. The Josh Bersin Company.
BuildEmpire (2025) Workplace Learning Habits Survey. London: BuildEmpire.
Gottfredson, C. and Mosher, B. (2011) Innovative Performance Support. New York: McGraw-Hill.
Merriam, S.B. and Bierema, L.L. (2014) Adult Learning: Linking Theory and Practice. San Francisco: Jossey-Bass.