I have watched more new capabilities die of low adoption than of low quality. The product was fine. Nobody used it.
Adoption is not downstream of the build; it is part of it. Incentives, air cover, talent design, and the path of least resistance for the user all decide whether a genuinely good tool ever gets picked up.
A capability nobody adopts is indistinguishable, on the balance sheet, from one that never existed.
Treat adoption as a first-class design constraint and the build changes shape. Treat it as someone else's problem and the best work quietly goes unused.