The first half earnings season saw an inflection in companies discussing artificial intelligence (AI), with around 15% of firms quantifying a revenue or cost-benefit from using machine learning across a broad spectrum of applications, Morgan Stanley (MS) said in a research report Monday.

The bank notes that 316 qualifying companies mentioned AI, with 106 explicitly citing fundamental business improvement from using AI or machine learning.

“29 quantified the revenue opportunity, 36 quantified a cost or productivity gain, 12 discussed customer service and in-housing as the lowest hanging fruit, 8 saw efficiency and in-housing of creative advertising, 82 other comments referred qualitatively to revenue or cost gains,” analysts led by Edward Stanley wrote.

The “lion’s share” of this activity occurred in the U.S., the note said.

The three most significant trends among non-tech company earnings were “biopharma gains from AI applications from strain selection to regulatory filings; large-cap banks discussing costs savings and onboarding from autonomous and in-house customer service; legal use cases from summaries to drafting appeared across the spectrum of reporting companies,” the bank said.

The bank’s analysis covered global companies with a market cap greater than $10 billion, which reported results since July 1.

Rival Wall Street giant Goldman Sachs (GS) said AI adoption will likely have a meaningful impact on the U.S. economy sometime between 2025 and 2030.

Read more: AI Will Be One of the Most Important Investment Themes Over the Next Decade: Morgan Stanley