AI confidence hides lack of savings, survey finds

WalkMe's survey paints a picture of a workforce that has moved from deception to self-assurance, but the confidence appears detached from measurable outcomes. CEO Dan Adika's anecdote about the Madrid conference—where only two people raised hands despite a survey saying 8% use AI meaningfully—sets the tone: actual adoption lags behind perception. The survey's drop in admitted deception might seem positive, but Adika argues it reflects comfort, not honesty: people no longer need to fake skill they believe they have. That belief is shaky, as half of workers report spending more time on AI tasks than doing them manually, and only a quarter say AI works on the first try.

The deeper risk, as Adika describes, is the 'company brain'—a system trained on employees' collective knowledge. The more thoroughly employees teach AI, the more replaceable they become, potentially shrinking teams from ten to three. This threat is most acute for Gen Z, who show the highest confidence and the most overstatement, with 31% reporting real workplace problems. Adika suggests the fix isn't more training, because overconfidence blocks learning; instead, he points to better integration and guidance within existing tools. His outlook is muted: without a mega breakthrough, he expects only marginal improvement. The story leaves a sobering note: workers may stop lying about AI, but they might start overestimating their own value in an AI-augmented workplace.