Ever since the introduction of ChatGPT, companies have been eagerly anticipating the day AI will turbocharge their workers and forever transform their businesses. Three years later, they’re still waiting. Why? And what’s the fix?
These two questions dominated pretty much every single conversation I had with executives in Davos last week — including a Business Insider roundtable I moderated with 15 chief people officers and other senior executives.
One explanation that came up again and again was incomplete adoption among employees. Many professionals are understandably worried about what these tools will mean for their jobs, or at least skeptical of their usefulness as AI slop abounds. To bulldoze through this hesitation, bosses have stepped up the pressure, making AI use mandatory and incorporating it into performance reviews.
But a number of executives at the roundtable advised against strong-arming. Cisco learned that the hard way. “When we asked our employees to take mandatory training for AI, not only did it not drive sustainable usage, it actually had a bit of a negative impact,” said Francine Katsoudas, the company’s chief people, policy, and purpose officer. What worked, she explained, was “providing choice” — like when Cisco gave its engineers access to half a dozen different AI tools, allowing them to decide which ones to use and how to use them. “They absolutely loved that,” she said.
Another theory was that even if employees want to use these new tools, they don’t have the necessary skills to get the most out of them. Part of the fix, some argued, involves hiring people who are already good at using AI. Kyle Lutnick, the executive vice chairman of Cantor Fitzgerald, said he wants to bring in more new college grads at a time when other businesses are hiring fewer of them, precisely because they have more fluency using these tools than their older counterparts. But hiring new blood won’t be enough. Employers will need to do a lot more to train their existing workforces too. “Investment has been primarily on the technology and not so much on the people,” said Elizabeth Faber, global chief people and purpose officer at Deloitte. “That needs to shift.”
A third explanation was that big productivity gains require a fundamental overhaul of the way work gets done inside companies. If the Googles or the Amazons of the world were to start from scratch today, they almost certainly wouldn’t have the team structures, workflows, and job descriptions they currently do. I think that’s why we’re seeing the AI revolution most clearly right now in early-stage startups, which are starting from the ground up in the post-ChatGPT era.
“84% of work processes have been left in their legacy state when adopting AI and have not been redesigned,” Faber said. “So 16% of organizations and work processes are really being developed in an AI-native way.”
All of these proposed solutions are far from quick fixes. Encouraging employees to opt in voluntarily takes more time than threatening to fire them. Training staff — and actually getting them to learn — takes time too.
Redesigning jobs will prove to be an even heavier lift. Many large businesses don’t even know what employees do on a day-to-day basis. It’s painstaking work to build out a comprehensive database of the skills employees have and the tasks they perform — and then to systematically tease out which of them can be delegated to AI and which of them can’t. One chief people officer I spoke to said that it’ll take years for her HR department to complete that process across every function at her company.
Once all that heavy-lifting is done, what will these businesses look like? I put the question to the group at the roundtable, asking how many of them expect their workforces to shrink in three to five years’ time. Two out of the 15 raised their hands — a tally I suspect would have been higher if I weren’t there. One of them, Gina Vargiu-Breuer, chief people officer at SAP, explained that her company is currently keeping headcount flat because the business is still growing.
“But when you’re not growing, then I think this is where you have to talk about, ‘OK, do we have to reduce headcount?'” she said. “I have a lot of peers in German companies where they are starting to reduce headcount dramatically. So it’s a reality. For us, it’s not, because we’re growing, but I think it will happen going forward.”
Even at SAP? I asked.
“At the moment we stay flat,” she said. “But if productivity goes up and growth is slowing down, then I think we have to look at that with different eyes.”
By the end of the week, I left Davos with a sense that a workplace truly reshaped by AI — one that would allow companies to run on meaningfully smaller teams — isn’t coming as soon as I’d thought. Those days seem to be still several years away, given all the painstaking work businesses need to complete to get there.
That’s something many economists had predicted early on, given how difficult it has always been to fully integrate new technology into the workplace. They told me that things will change less than we expect in the short term, and a lot more in the long run. No matter how fast a technology advances, humans change less readily.
C-suites around the world are coming to the same realization, which is probably why I detected quite a bit of frustration in Davos. Svenja Gudell, Indeed’s chief economist, compared the world’s urgency around AI to the impatience of a parent potty-training their kid. “It’s messy, it’s a long process,” she said. “You’re like, ‘Why is this not happening? It’s been three weeks already.'” Her message to executives: “Give yourself some grace.”
The slower timeline is good news for the rest of us — it gives us time to learn new skills, debate new public policies, and try to shape the future we actually want. But it would be a mistake to read the so-far modest changes as evidence that tectonic shifts aren’t coming. I came away from Davos convinced that when they do happen, they’ll be far bigger than anything we’re imagining now, for better and for worse.
Aki Ito is a chief correspondent at Business Insider.
// Indeed
const seriesTitle = “”;
// Presented by
const text = “Presented by”;
// 6977990fe1ba468a96aab2e2?width=1441
const sponsorLogoID = “6977990fe1ba468a96aab2e2?width=1441”;
// Indeed
const altText = “Indeed”;
//
const hubOrCatURL = “”;
document.documentElement.classList.add(“gi-sponsor-module”);
// Selectors and styles last updated: 6-5-25
if (
document.querySelector(“.gi-sponsor-module”) &&
document.querySelector(“.post-body”) &&
!document.querySelector(“.is-enhanced”)
) {
document.querySelector(“.post-hero”).insertAdjacentHTML(
“afterend”,
`
${text}
`
);
}
if (
document.querySelector(“.gi-sponsor-module”) &&
(document.querySelector(“figure.as-full-bleed”) ||
document.querySelector(“.overlay-hero-image”)) &&
document.querySelector(“.is-enhanced”)
) {
document.querySelector(“.timestamp”).insertAdjacentHTML(
“beforebegin”,
`
${text}
`
);
}
if (
document.querySelector(“.gi-sponsor-module”) &&
document.querySelector(“.breadcrumbs-wrapper”)
) {
document.querySelector(
“.breadcrumbs-wrapper”
).innerHTML = ``;
}
Read the original article on Business Insider
The post Executives are coming to terms with AI’s slow impact on work appeared first on Business Insider.
