AI dashboards miss the real problem at work

Companies are racing to measure AI adoption through dashboards, token counts, platform engagement, and even employee leaderboards. Some have gone further, tying AI use to raises and promotions. But according to the author, Cornerstone’s Chief AI Officer, that focus is creating a misleading picture: usage is not the same as capability.

Research based on a survey of 2,000 workers across the U.S. and U.K. points to the gap. While 46% of employees say they use AI tools at work, nearly half have received no formal AI training, and 56% have no clear path to building AI-related skills. More strikingly, 17% say they are pretending to use AI at work.

That matters because activity metrics can look healthy while real competence remains weak. An employee producing 10 prompts a day may seem engaged, but that does not show whether they can write effective prompts, check outputs, spot hallucinations, or use AI responsibly in ways that improve performance.

Why AI usage metrics fall short

The piece argues that many organizations are moving faster on measurement than on workforce development. As AI becomes part of everyday work, leaders want visible proof that their investments are paying off. Dashboards and prompt counts offer that visibility, but they can also create false confidence.

Recommended reading

6G signal breakthrough filters interference at the source

The same problem appears in process automation. The author warns against trying to “agentify” everything by wrapping AI agents around existing workflows without rethinking the work itself. Automating a broken process, they argue, only creates a faster broken process. And measuring success only through efficiency gains undersells the opportunity: the bigger prize is transformation that improves both topline and bottom-line results.

CIO and HR need joint accountability

A core problem is fragmented visibility. One team manages technology procurement, another owns learning and skills data, and performance data often sits elsewhere. That makes it hard to connect AI usage with actual business outcomes.

The author says this is as much a CIO problem as an HR one, and calls for joint accountability rather than simple coordination. In that model, CIOs and HR leaders would share responsibility for whether employees can actually execute the company’s AI strategy.

The piece also argues that the skills most likely to endure are not tied to any single model or tool, but to domain expertise expressed as work. Organizations that connect AI adoption directly to real tasks and outcomes are more likely to build durable capability instead of superficial usage.

TechRadar notes the article was published as part of TechRadar Pro Perspectives and that the views are those of the author, not necessarily TechRadarPro or Future plc.