⚠ AI Flattens the CurveLow threat

Learned-skill lock-in assumes that fluency in a tool is hard-won and costly to replace — but that assumption is exactly what artificial intelligence threatens to undo. If an AI assistant can let anyone accomplish in an unfamiliar program what previously took years of practice to master — writing the formula, building the report, formatting the document simply by asking — then the trained-in skill that bound the workforce to Microsoft's tools becomes far less of a barrier to using someone else's.

The danger strikes at the mechanism directly. The whole point of learned-skill lock-in was that switching meant retraining a hundred thousand people at ruinous cost; but if AI flattens the learning curve of any tool to near-zero, the retraining cost that made migration unthinkable collapses, and a rival suite with its own AI helper becomes usable on day one. The moat depended on the difficulty of learning, and AI is in the business of making things easy to learn.

Microsoft's hedge is that it is the one putting the most capable AI assistant inside its own tools, so if AI makes any suite instantly usable, it makes Microsoft's — already the incumbent — instantly usable too, giving customers even less reason to leave. And skill was only one thread; the compatibility network and the deep integrations do not care whether a newcomer can learn a rival tool quickly.

Hold it at low-to-moderate. This is a genuine long-run corrosion of one specific lock-in — the trained-workforce barrier — and it deserves watching precisely because Microsoft's own technology drives it. But skill was never the only thread binding the customer, and Microsoft's incumbency lets it turn the same AI that flattens the curve into one more reason to stay — it shipped Copilot into Office before rivals shipped anything comparable at all¹.