Executives are describing AI’s impact on productiveness largely as one thing nonetheless to return, in response to new analysis from the Federal Reserve Financial institution of St. Louis, at the same time as mixture productiveness reveals little or no precise productiveness positive factors.
What are execs saying about AI?
The St. Louis Fed analyzed about 490,000 earnings name transcripts from 5,198 publicly traded U.S. corporations overlaying 2000 via 2025. Researchers used an open-source massive language mannequin to flag each sentence about productiveness, decide whether or not it additionally talked about AI and classify whether or not the speaker was describing a previous, current or future acquire. The share of productivity-related sentences that additionally talked about AI was close to zero earlier than ChatGPT’s launch, climbed via 2023, plateaued in 2024, then accelerated once more in 2025 to about 15% of productiveness sentences by year-end.
Throughout AI-related productiveness sentences, about 95% referred to future positive factors, in contrast with roughly three-fourths of non-AI productiveness sentences. Tone adopted the identical sample: 95% of AI-related sentences described productiveness as rising, in opposition to 75% for non-AI sentences. Fed researchers Serdar Ozkan, Aakash Kalyani and Nicholas Sullivan write that companies are “investing, experimenting and reorganizing round AI at the moment, whereas the measurable productiveness results stay principally forward.”
Who’s most optimistic about AI productiveness?
That optimism has not but proven up within the broader financial system. The researchers cite Federal Reserve Financial institution of San Francisco knowledge displaying utilization-adjusted whole issue productiveness grew simply 0.07% over the 4 quarters ending within the first quarter of 2026. A associated San Francisco Fed examine by Kalyani and Huiyu Li discovered that companies with optimistic AI sentiment on earnings calls have posted greater capital spending and analysis and growth progress than different public corporations. They notice this combo of excessive optimistic sentiment and excessive spending is concentrated amongst massive expertise companies constructing AI infrastructure.
The researchers say they’ll proceed monitoring whether or not the language in earnings calls shifts from expectations to realized ends in the approaching quarters. For HR leaders, the analysis presents a helpful counterweight to inner narratives that deal with AI productiveness positive factors as realized. Workforce plans, hiring freezes or layoffs constructed on the belief that AI has already delivered effectivity positive factors are operating forward of what the information present. Executives on earnings calls are themselves describing these positive factors sooner or later tense to traders.
