Only 4% of US HR chiefs feel very prepared for AI-driven workforce change

Workers are using AI more every day. Fewer of them can get training than a year ago.

PwC surveyed 49,364 workers across 48 countries and regions for its 2026 Global Workforce Hopes and Fears Survey. Between 2025 and 2026, the survey found, access to learning and development dropped from 59% of employees to 51%. Writing in Human Resources Director America, Jeffrey R. Smith calls this a design failure, not a technology failure. Organizations are rolling out new tools while leaving unchanged the work those tools were supposed to support.

The Conference Board has measured how ready HR leaders feel. Its CHRO Confidence Index for the third quarter of 2026 found that just 4% of US chief human resources officers describe their organizations as very prepared for AI-driven workforce change in the coming two years. A little over half say they're prepared to some degree.

The hard question is whether anyone has looked closely at the jobs that AI licences land in.

Substitution versus reinvention

Ram Srinivasan, a managing director at JLL in Chicago, calls the split substitution versus reinvention. JLL surveyed more than 2,200 C-suite executives worldwide for its 2026 AI research. Srinivasan says most organizations are still substituting, applying AI to processes they already run. The companies producing measurable returns have moved on to asking whether those processes should exist at all.

That helps explain how usage numbers can look healthy while readiness stays low. Reinvention appears as a changed workflow, a handoff removed, a role rewritten.

Skillsoft's Workforce Readiness Report on AI found that training often arrives after tools are released. According to Skillsoft, 86% of individual contributors and 93% of managers now use AI at work, many without training, organizational guidance or confidence in their long-term skills. Just 16% of individual contributors and 23% of managers report receiving training before new AI tools roll out. That suggests many companies are deploying AI before they prepare employees to use it.

The perception numbers are the most revealing. In the same report, 77% of managers believe employees are set up for success with AI, but only 24% of individual contributors strongly agree that their employer has prepared them. Skillsoft calls this a clear gap between leadership confidence and employee experience. The questions differ slightly, since managers were asked about their teams and contributors about themselves, but the direction is clear.

Skillsoft found manager evaluation is the main way companies judge what people can do, cited by roughly a quarter of individual contributors and 29% of leaders. Skillsoft says the perception gap suggests organizations may be overestimating readiness. For companies that depend mostly on manager evaluation, that risk could be higher.

Vistek's decade of process maps

At Vistek, a Canadian imaging and technology retailer with about 160 employees, Lee-Ann Reid is both vice president of operations and director of HR. Before she looked at where automation could add value, she told HRD America, she put six to 10 years into mapping the process behind every workflow. That's a long runway. It's also why her AI work had a foundation to build on.

Joey Price, writing in HR Executive, separates chatbots that answer questions from AI agents that can coordinate workflows across enterprise systems.

Reid also argues that HR professionals are trained to set aside emotional attachment to a role and look at the problem itself. That's useful when a map shows a task nobody needs to do anymore.

HR and IT on the same rollout

Fiona Ho is CHRO at Sophos and works from Vancouver. Sophos is a global cybersecurity company headquartered in Abingdon, U.K. Ho urges HR leaders to work closely with IT on AI rollouts. Sophos treats its rollout as a workforce transformation that pulls several functions together across the whole company. She doesn't see it as a technology project that HR gets briefed on afterward.

In Ho's view, tool access is only half of effective AI adoption. The other half is building basic human skills: thinking critically, communicating clearly and taking responsibility for what AI produces. Handing an employee a licence, she says, does not hand them a capability. She calls the distance between those two things a design gap rather than a training gap, and says closing it takes a different kind of investment. Skillsoft's survey found that 17% of individual contributors completely trust what AI gives them, and about 14% say they are cautious and verify outputs before using them. The most common answer, from 30% of individual contributors, was that they trust AI in most situations, and a quarter trust it for specific use cases.

Pascal Gouel is managing director of TAG, a Sydney-based casual labour hire firm that manages approximately 5,000 workers for close to 100 blue-chip clients in aviation, hospitality, fast-moving consumer goods, supply chain and retail in Australia. As AI takes over information-gathering, analysis and routine execution, he says, technical AI skill is becoming less scarce. What is gaining value is sound decision-making within a governance framework, and he means at every level of the organization, not just in the executive suite.

In Skillsoft's survey, only 9% of individual contributors and 12% of leaders reported company-wide governance for AI tools. For employees expected to make decisions within a governance framework, that suggests many could be working without one.

The case against waiting for the map

Time is the strongest objection to a design-first approach. Reid says she spent six to 10 years building process maps for every workflow at Vistek, a retailer with approximately 160 employees.

Price, whose firm Jumpstart HR runs the Insights@Work study, gives that objection weight. His 2025 edition surveyed hundreds of employers. Only 6% had implemented AI broadly across functions, 31% were piloting or testing, and more than a third were still working out what they wanted AI to do. 42% said employees lacked the skills to use it well. Price wrote that technology developers are discussing intelligent agents, workflow orchestration and autonomous systems, while many HR practitioners are still debating whether AI is appropriate for their organizations. He sees the bigger risk as HR's grasp of the technology falling behind the work it's responsible for leading.

Price notes that product demos assume connected systems and clear workflows, which real organizations rarely have.

KPMG's bet and the next readings

Jason LaRue, vice chair of talent and culture at KPMG U.S., has described the challenge as a "work trifecta." First, the work itself is being redesigned. Second, the workforce is changing in which capabilities get hired and rewarded. Third, the workplace, both physical and digital, is evolving in how people collaborate. According to LaRue, KPMG's investment in simulation-based learning has aimed to develop uniquely human skills, including judgment and critical thinking. Those are the same skills Ho and Gouel point to, and a licence doesn't supply them.

The Conference Board's very-prepared share stands at 4%. PwC's learning-and-development access figure stands at 51%, down from 59% in 2025.