Companies still don't quite know what to do with AI

Behind the scenes: The key question now is: Can widespread individual adoption of AI actually translate into meaningful gains for an organization? At a recent private dinner Axios attended, HR executives debated the issue.

• One participant described inheriting a $10 million AI bill without a clear explanation of the value created.
• Entire workflows might need a redesign to really capture the opportunity on offer, other executives said.
• "Productivity is ultimately shaped by how people work together," notes a summary report of the evening, hosted by Charter.

Zoom in: "Almost every company's had this experience," says Brian Elliott, who is executive in residence at the workplace consulting firm.

• "Individuals report gains in productivity, but they're not seeing any results really from a company or or even parts of an organization perspective."
• AI has made writing code faster, he says, but hasn't sped up the process of figuring out what to do with it.

What to watch: CEOs are under increasing pressure from the boardroom to actually wring some value out of all their AI spend.

• And they're starting to push out executives they think aren't moving the needle, says Jeff Christian, CEO of executive recruiting firm Christian and Timbers.
• "We're seeing some companies replace 20 to 30% of their executive leadership team that can't make the jump into AI and drive the transformation that needs to occur."

What they're saying: "Some of the simplistic early views are that you just sort of somehow put technology inside of an organization and it creates a transformation," Rob Hornby, global co-CEO of AlixPartners, tells Axios. "But that's never happened and it's not happening in AI either."

• He says the consulting firm is seeing an influx of business as companies try and figure out what to do with AI.
• "Almost every company in the world needs to do a substantive transformation. So it's a massive creator of demand."

Friction point: Companies are also holding off on major transformations out of security concerns. Banks worried about hacks — or being on the hook when a rogue agent uses a customer's account to buy something the customer doesn't want — for example.

• "Anyone who thinks they might be liable further down the chain is just stopping the activity," Hornby says.

State of play: It helps explain why AI isn't yet clearly showing up in the overall productivity numbers, as Axios' Courtenay Brown recently reported.

• In Anthropic's filing to go public, the frontier AI company called out the problem: "The natural resistance to change within organizations" is an obstacle to its growth.

Between the lines: Technological transformations take time —  it took decades for electricity's benefits to transform factories.

• Seems hard to believe now. The increase to productivity from computers was slow, as well, prompting, Nobel-winning economist Robert Solow's famous quip in 1987: "You can see the computer age everywhere but in the productivity statistics."
• The situation now with AI is sort of akin to inventing cars without yet building the roads or highways to drive on, Ben Dattner, an organizational development consultant, tells Axios.

Where it stands: Not seeing a payback yet, companies outside the tech sector are pulling back, Apollo's chief economist, Torsten Slok, wrote recently.

• "Despite massive investments and widespread hype around artificial intelligence, the technology has delivered zero financial payoff to corporate America outside of tech."
• Revenue per worker and profit margins have been flat for the S&P 500 companies outside the Magnificent Seven tech firms, per data Slok shared.

Yes, but: In software engineering, there's little question AI had altered work and workflows.

The bottom line: AI may be moving fast, but companies are afraid to break things.