Employees quietly using AI while teams have spare capacity

Prodoscore has published workforce data showing unused employee capacity and widespread use of consumer AI tools outside company oversight.

The analysis covered activity across 72 companies and highlights two issues many employers may not be tracking closely: how much spare capacity exists within teams, and how often staff use AI products outside formal IT controls.

The data shows 23.7% of employees had materially more unused capacity than peers in similar roles. That was little changed from 22.7% a year earlier, suggesting a steady pattern rather than a short-term fluctuation.

By contrast, almost no employees fell into the equivalent group for over-utilisation. That imbalance suggests organisations may have more room to absorb work internally than workforce planning models often assume.

Prodoscore measures visible on-screen activity and uses it as a signal of capacity, not a judgment on an individual's productivity or working hours. It says the figures should be read alongside output, duties and performance rather than used as a stand-alone measure.

Capacity question

The findings come as employers continue to assess staffing levels after years of changes in hybrid work, digital monitoring and shifting expectations around efficiency. Companies that maintain hiring plans without a clear view of whether existing teams are fully occupied risk adding headcount where work could be redistributed.

That does not necessarily mean businesses can simply cut recruitment. Spare capacity in one role or team does not always transfer neatly to another, and activity levels alone do not show whether workers have the right skills, seniority or business context to take on more tasks.

Still, the absence of a meaningful over-utilised group stands out. It suggests the bigger issue in the sampled companies was not widespread employee overload, but limited visibility into where time may be underused.

Sam Naficy, Chief Executive Officer of Prodoscore, said leaders should look internally before deciding to hire. "These findings suggest that leaders should first evaluate the capacity that already exists within their teams before automatically opening a new position," he said.

Naficy said activity data should not be viewed in isolation. "When considered alongside employee output, responsibilities and demonstrated performance, it can help companies make more informed workforce decisions. Prodoscore measures visible, on-screen activity and presents it as a capacity signal, not a verdict on individual productivity or number of hours worked," he said.

Shadow AI

A separate part of the analysis examined AI use within the same group of companies. Employees actively used nearly 50 different AI tools, pointing to a fragmented market in which workers appear willing to experiment beyond the software formally provided by employers.

Three of the five most-used tools were consumer products: ChatGPT, Gemini and Claude. Their prominence suggests employee behaviour is moving faster than many corporate procurement and governance processes.

ChatGPT was the clearest example. It reached nearly three times as many employees as Microsoft Copilot and accounted for nearly five times as many usage hours, according to the data.

That gap matters because consumer AI tools often operate outside managed company environments. In practice, employers may have limited knowledge of which systems workers are using, how often they rely on them and what information they enter into prompts.

For technology and risk teams, the issue is not only whether staff use AI, but whether policy has kept pace with use already embedded in day-to-day work. A company may have approved one assistant for official use while employees quietly turn to several others because they prefer the interface, responses or availability.

The data also suggests formal deployment alone does not guarantee adoption. Even where an employer has introduced an approved product such as Copilot, workers may still gravitate to consumer alternatives if they find them easier to access or more useful for specific tasks.

Naficy said the two trends should be viewed together. "Taken together, these findings reveal two significant blind spots for today's employers.

"Organisations may be underestimating the capacity that already exists within their workforce while simultaneously underestimating how quickly employees are adopting AI outside of company systems. Closing both visibility gaps will help leaders make smarter hiring, workforce planning and technology investment decisions," he said.