Only 5% of staff use generative AI effectively, study says

NROC Security has published research showing that about 5% of employees using generative AI apps qualify as effective users. The study also found that 31% of eligible employees actively use the tools.

The findings are based on a sample of 4,800 business users and 139,000 generative AI interactions across what NROC described as a typical enterprise mix of knowledge-worker profiles.

The data points to a widening gap between access to generative AI and the ability to use it in ways that deliver measurable workplace gains. Active usage nearly doubled from NROC's previous quarterly study, rising from 19% to 31% of eligible employees, while the number of app families in use increased from 19 to 53.

Yet the share of workers classed as effective users remained small. Under NROC's methodology, that group consists of daily or weekly users who also show a high level of prompting skill.

This segment had grown tenfold from the earlier study, but still represented only about 5% of eligible employees. The contrast suggests that broader rollout of AI tools inside businesses has not yet translated into widespread effective use.

Usage gap

NROC's report argues that frequency of use and prompting skill are stronger indicators of productivity than simple access to AI applications. That is the central conclusion of its latest quarterly analysis.

"We are encouraging our employees to use GenAI apps, but are our businesses becoming more productive?" asked Antti Reijonen, Chief Executive Officer of NROC Security.

He said the question had become common among leadership teams as adoption of consumer and workplace AI tools spreads across organisations.

"It's one of the most common questions we hear from leadership teams, and one of the hardest to answer. Our second quarterly study looks into it," Reijonen said.

The report also tracks how employees are using specific app families, including ChatGPT, Claude and Microsoft Copilot, alongside prompting-skill data and data-exposure trends. It also sets out the methodology behind what NROC calls a GenAI Effectiveness Index.

Governance focus

Alongside the usage benchmarks, NROC is promoting a framework it describes as productivity-first governance. The approach centres on allowing broad access to tools while maintaining oversight, supported by guardrails and team-based efforts to improve prompting skill.

"Our Q2 report's core finding is that frequency and prompting skill, not access to AI, are what actually drive productivity," Reijonen said.

He added that simply issuing licences across an organisation may have little effect if employee use remains infrequent or unsophisticated.

"An organization can roll out enterprise licenses to everyone and still see almost no effectiveness gain if usage stays occasional and prompts stay basic.

"This new report lays out a practical framework for closing that gap - productivity-first governance - built around allowing broad access while maintaining visibility, using guardrails as a safety net rather than a wall, surfacing what top users already do well, and building prompting skill team by team," he said.

The research enters a market debate that has shifted from early adoption rates to the value companies are getting from generative AI tools. Many organisations now face pressure to show whether spending on licences and internal AI programmes is changing how staff work.

For technology buyers and security teams, the figures also underline how difficult it can be to separate experimentation from sustained, effective use. A workforce may show rising activity across multiple AI tools without a similarly broad increase in productivity.

NROC focuses on governing employee use of generative AI applications. It sells a network-based AI security proxy intended to help organisations oversee access to popular tools, monitor prompts, responses and file uploads, and support compliance with internal policies.

The rapid spread of generative AI in the workplace has created fresh concerns around data privacy, policy enforcement and visibility. NROC's research suggests those governance questions now sit alongside a second challenge for employers: how to move workers from occasional use to repeatable, effective use.

Renaissance Distribution is NROC's distribution partner in the UK and Ireland. The distributor works with cybersecurity, cloud and infrastructure vendors through channel partners including managed service providers, resellers and systems integrators.

NROC plans to continue publishing quarterly data as the basis for its effectiveness index. The latest results show that while generative AI use is moving into the mainstream of office work, the share of employees using it well enough to make a meaningful difference remains limited at about 5%.