AI isn’t enough to drive innovation—HR needs to focus on talent strategy

Would you be surprised to know that AI use isn’t moving the needle for most organizations?

The majority (88%) of global employees report using AI at work, but just 28% of companies have seen a discernible impact on “business transformation,” a recent report from consulting firm EY found. This is the case, despite the fact that 50% of leaders, 42% of knowledge workers, and 33% of essential workers report using the tech daily.

AI adoption and use is accelerating at a time when some companies have “fragile types of talent dynamics,” said Roselyn Feinsod, a principal in EY’s people consulting practice. This, she said, could be hampering innovation and transformation. As AI use increases, she recommended HR focus on improving talent strategies.

“The magic formula that we saw that really kind of gets you to the strongest point is built on the mix of what we are describing as skillset, toolset, and mindset as it relates to AI,” Feinsod told HR Brew.

What does this mean for HR? A skillset can be achieved through L&D, specifically AI upskilling and reskilling, Feinsod said, and a mindset can come from people leaders setting employee goals around AI use. A toolset, meanwhile, relies on employees to use company-endorsed AI tools, on top of their own personal AI tools, she said. This is also known as “shadow AI use.”

“The world is moving so fast that we’re not just relying exclusively on the solutions that our companies are providing to us, but also playing with our own personal versions of what we are seeing as a best fit for our job and role,” Feinsod said.

While HR has a responsibility to put guardrails in place to protect sensitive information from being shared with AI tools, Feinsod said it’s important to listen to employees about the tech they’re using. If it helps them do their jobs better, she recommended HR pros advocate for it to be incorporated into their suite of company-endorsed AI tools.

“It just is going to keep going up every year in terms of the frequency and the complexity of the different types of ways that employees can use these things,” she said. “Also, as things get all the way to agentic, where it’s not even the employee using it, but the employee reviewing work of agents who are going to be doing some of that work for us.”