The usual first step in AI upskilling is a training calendar: pick the courses, book the rooms, count the attendees. The step that should come before it is an honest inventory of what the team can already do. Skipping it means paying to teach people things they know, while the things they do not know go unaddressed.
Much of what makes a team good is tacit. It is the judgment a senior coordinator uses to spot a supplier delay before it is reported, or the instinct a reviewer has for which line in a contract will cause trouble. Nobody wrote it down, so no competency framework lists it and no training-needs survey asks about it. Recognition of prior learning exists to deal with this: it assesses competence people have built through work rather than in a classroom, and treats evidence of doing the job as evidence of capability. A pilot of this approach with SkillsFuture Singapore, building a framework for recognising workplace-acquired competency at scale, is where the S.E.N.S.E. framework's starting premise comes from.
Applied to AI, the inventory becomes a capability gap map with three columns. In the first are capabilities the team already has and can demonstrate. In the second are capabilities that exist but live in one person's head. In the third are capabilities that are genuinely absent. Training money belongs mostly in the third column. The second column needs documentation and a way to share it. The first needs protecting, because it is exactly what a careless workflow redesign breaks.
That last point deserves attention. As AI absorbs routine tasks, the capability that stays valuable is judgment: noticing that an output is wrong, knowing what a client meant rather than what they wrote. These are the capabilities least likely to appear on a skills matrix, so they are the most likely to be reorganised away without anyone deciding to.
There is a fairness argument as well. Decisions about people, such as who is redeployed, whose role is redesigned or who is offered retraining, should rest on evidence of what they can actually do, not on a job title. An evidence-based inventory gives people a fair look before those decisions are made.
Once the map exists, the learning plan for the first quarter writes itself, because it is sequenced to what that specific team is missing. A simple test before approving the next training budget: can you list five things your team does well that appear in no job description?
