DAVOS, SWITZERLAND — When it comes to the economic impact of artificial intelligence, is 2024 going to be more like 1987 or more like 1995?

Driving the news: That, in a nutshell, is the question beneath much of the (abundant) AI discussions taking place among global leaders and top thinkers at the World Economic Forum this year.

Why it matters: In the 1990s and early 2000s, a revolution in information technology helped fuel a productivity boom — and with it, an environment of rapid growth, rising wages and low inflation.

• Many in the Davos crowd envision something similar — or more significant — emerging from AI advances. Less clear is when.
• Yet the 1990s IT productivity boom didn't arrive until late in that decade, even though it was based on technologies developed in the 1980s. "You can see the computer age everywhere but in the productivity statistics," economist Robert Solow famously observed in 1987.

State of play: It takes time for companies to learn how to use technological innovations to their maximum effect to get more output from their workers. In Davos, many talks have centered on AI's leap from an interesting novelty to the core driver of business efficiency.

What they're saying: "We're at a phase where in 2024 we think generative AI will move from pilots and experiments to implementation and industrialization," Paul Knopp, the U.S. CEO of KPMG, tells Axios.

• Clara Shih, the CEO of Salesforce AI, said at Axios House on Monday that the companies Salesforce works with "are already seeing productivity gains" from AI tools.
• In sales, software development, marketing and customer service, "when you look at the metrics that customer service leaders typically run their operations by: average handle time, customer satisfaction, cross-sell ... all of those metrics are rising."

After all, implementing AI-driven work processes doesn't require the same kind of massive investment in equipment that earlier waves of technology have; most knowledge workers already sit at a computer all day.

• "I'm actually very bullish about some of the possibilities for significantly improving productivity much faster than we might have thought," said Stanford AI economics scholar Erik Brynjolfsson at another Axios House event.

Yes, but: Companies usually don't rework their processes overnight. Software must be vetted for security and accuracy. Employees need to be retrained. That may paradoxically slow productivity growth during implementation.

• "AI is foundational technology, like electricity," Karen Harris, managing director of Bain & Company's Macro Trends Group, tells Axios. "It's so nascent now, we don't know how quickly we'll see the impact at a macro level."
• "There are businesses that see the disruption already. But will that aggregate up?"

The bottom line: 2024 will be a year of significant change thanks to AI, but there is no certainty it will be visible, as Solow once said, in the productivity statistics.