Executives at publicly traded U.S. companies increasingly link artificial intelligence to productivity improvements, but nearly all of that talk remains focused on future expectations rather than realized gains, according to an analysis published in July 2026 by the Federal Reserve Bank of St. Louis. Researchers examined roughly 490,000 earnings call transcripts from 5,198 firms spanning 2000 through 2025, finding that approximately 95% of productivity-related sentences mentioning AI referred to future improvements. The report highlights a sharp disconnect: while executives express overwhelming optimism about AI's potential, aggregate productivity data through the first quarter of 2026 showed utilization-adjusted total factor productivity grew just 0.07%.
The share of productivity-related discussion devoted to AI climbed dramatically after ChatGPT's late 2022 debut. By the end of 2025, roughly 15% of all productivity sentences in earnings calls also mentioned artificial intelligence, up from near zero beforehand. The analysis also revealed that overall productivity discussion rebounded after a sharp pandemic-era dip, with productivity-related sentences accounting for a larger portion of transcripts by 2025 than before COVID-19. Language patterns shifted as well: in 2023, "generative AI" dominated, but by 2025, phrases like "using AI" and "AI tools" became more common, suggesting firms moved from describing the technology in abstract terms to discussing practical applications.
When executives talked about AI and productivity together, sentiment was overwhelmingly positive. The report found that 95% of AI-related productivity sentences described productivity as increasing, with negligible mentions of declines. Even non-AI productivity discussions skewed optimistic, with 75% describing rising productivity across all quarters from 2000 through 2025. The forward-looking nature of AI discussions stood out: approximately three-fourths of non-AI productivity sentences referred to future gains, but that figure jumped to 95% when AI entered the conversation. This pattern held consistent in both 2023 and 2025, indicating that even as time passed after ChatGPT's release, executives continued describing AI's productivity effects as something yet to come.
The researchers used a large language model to extract and classify sentences from transcripts, identifying whether discussions involved AI, whether productivity was described as rising or falling, and whether statements referred to past events or future expectations. This approach captured roughly 910,955 tagged productivity sentences, each linked to a specific firm and quarter. The report notes that firms may currently be investing in and reorganizing around artificial intelligence while measurable productivity effects remain mostly ahead. A separate 2026 San Francisco Fed study cited in the report showed that firms expressing positive AI sentiment substantially increased investment growth by 2025, with recent capital spending and research and development growth driven largely by AI-positive companies, especially major technology firms building AI infrastructure.
The analysis suggests artificial intelligence has already reshaped how companies discuss productivity, even though its effects haven't clearly appeared in aggregate productivity statistics. The authors write that earnings calls currently point less to broad realized productivity gains than to a corporate sector actively investing in, experimenting with, and expecting future gains from AI. The researchers plan to continue monitoring whether AI-related productivity discussion in upcoming earnings calls shifts from future-tense expectations to reports of realized improvements, offering a window into whether executive optimism eventually translates into measurable economic gains.