Overview

• The study, based on interviews, surveys and more than 300 U.S. projects, finds only about 5% boosted sales, a release that coincided with a late‑August pullback in AI‑linked stocks and revived bubble concerns.
• Researchers describe a learning gap inside firms, pointing to failures in adapting processes, data, memory and governance rather than tool capability.
• More than half of generative‑AI budgets go to sales and marketing despite higher returns typically coming from internal automation, reduced outsourcing and operational efficiency.
• Employee use of tools such as ChatGPT is advancing through a shadow AI economy that often outperforms formal, large‑scale deployments at the company level.
• Investors are pressing for clearer disclosure on implementation and oversight, while in Argentina many firms invest under 1% in analytics/AI and give genAI low strategic priority despite strong public demand for practical training.