A company runs a training program. The right people attend, the slides land, the certificates go out. Six months later, almost nothing has changed.
It isn’t the training’s fault. It’s the format. A classroom sits at a distance from the actual work, and the lesson rarely survives the walk back to the desk. By the time there’s a real chance to use it, the moment and most of the material are already gone.
McKinsey’s Suman Thareja put it plainly: “A typical learning journey doesn’t cut it anymore. Learning has to happen in the flow of work, cross-functional, hands on, and together.” That one sentence captures what most L&D programs miss. Capability isn’t built in a classroom. It’s built in the work itself, when someone takes on a stretch project they’ve never done before alongside a mentor who has, or when a junior PM sits in a meeting with a senior engineer they’d normally never be in the room with. Yesterday’s lesson becomes today’s project, and last quarter’s gap becomes this quarter’s capability.
The stakes have risen, too. In the AI era, every employee is being asked to become a technology-enabled problem solver, not just the engineers and not just the executives. That is a far bigger shift than handing people new tools and hoping they figure it out.
It’s also where two kinds of companies begin to separate. Some give their employees access to AI and call it progress. Others redesign how the work itself gets done: they embed learning into daily work, shorten the gap between learning something and using it, and build the kind of continuous-learning system that actually holds up in practice. The difference between those two companies shows up, before long, in performance.
One global organization recently made that shift. It moved from a role-based model to a skills-based one, redesigned its career paths, built skills into performance management, and planned its workforce around hiring, reskilling, and redeployment. The unlock was visibility: a clear view of more than 2,000 critical roles and the skill levels behind them.
But visibility was only the starting point. Once a system can see who has which skills, who is quietly drifting, and who is ready for more than they’ve been given, it can finally do the work companies have always struggled to do by hand. It pairs mentors with the people they’re best positioned to teach and routes stretch projects to those quietly ready for them. It maps career paths before someone outgrows their seat, surfaces replacement candidates before a critical role goes empty, and connects people across teams who would otherwise never meet.
That is what learning in the flow of work really looks like. It isn’t better training. It’s a system that connects the moving pieces in real time.
None of it happens on its own, though. McKinsey’s Tanguy Catlin points out that the hardest part of any transformation isn’t the strategy or the technology, it’s the change management: aligning the organization and making the value clear enough that people actually come along. That alignment starts with seeing your people clearly, and then doing something about what you see.
The advantage in the AI era won’t go to the companies with the best tools. It will go to the ones that can see their people clearly enough to match them to the work that grows them, every day, in the flow of the work itself, instead of waiting for the next training session.
Training shaped last decade’s teams. The work shapes the next.
Source
- McKinsey & Company, Two McKinsey partners on rewiring talent for the AI era — Suman Thareja on learning in the flow of work, Tanguy Catlin on change management as the hardest part of any transformation, and the global organization that mapped 2,000+ critical roles. mckinsey.com