The state of enterprise AI 2025

OpenAI's "The State of Enterprise AI 2025" report presents an analysis based on data from over one million enterprise customers and surveys of 9,000 workers across nearly 100 companies. Aimed at business leaders, technology directors, and digital transformation professionals, the document examines how organizations are integrating AI into their daily operations and what results they are achieving.

Adoption growth is exponential. Message volume in enterprise ChatGPT has increased 8-fold in one year, while reasoning token consumption has multiplied by 320. Tools like custom GPTs and Projects have grown 19-fold, processing 20% of all enterprise messages. This demonstrates that AI is moving from one-off experiments to becoming integrated into structured and repeatable workflows.

The productivity impact is significant. 75% of workers report improvements in speed or quality, saving between 40 and 60 minutes daily. Technical roles like engineers save between 60 and 80 minutes. Additionally, 75% claim they can now perform tasks they couldn't before, such as data analysis or programming, democratizing technical capabilities. Code-related messages have grown 36% in non-technical teams.

The document includes real cases with tangible results. Intercom resolved 53% of calls without human intervention, Oscar Health resolves 58% of medical queries instantly, BBVA automated 9,000 annual legal consultations, and Moderna reduced pharmaceutical document analysis from weeks to hours. These cases demonstrate improvements in efficiency, customer experience, and innovation acceleration.

Adoption is global and multi-sector. Technology, healthcare, and manufacturing show the fastest growth, while finance and professional services operate at greater scale. Although the United States leads in volume, Australia, Brazil, France, and the Netherlands experience growth rates exceeding 140% annually.

A critical finding is the growing gap between leaders and laggards. Frontier workers send 6 times more messages than average, and leading companies use 7 times more custom GPTs. Users who apply AI across seven types of tasks save five times more time than those who use it in only four.

The report concludes that the main obstacle is no longer technology, but organizational readiness. Successful companies invest in infrastructure, governance, leadership, and change management. AI is ceasing to be an individual productivity tool and becoming central infrastructure for modern enterprises.

Key points

• Data from over 1 million enterprise customers and 9,000 workers show accelerated AI adoption
• 8x growth in ChatGPT message volume and 320x in reasoning tokens per organization
• 75% of workers report improvements in speed or quality, saving 40-60 minutes daily
• Democratization of technical tasks: 75% can perform activities they couldn't before like data analysis or programming
• Real cases demonstrate impact: process automation, response time reduction, innovation acceleration
• Global multi-sector growth: technology, healthcare, and manufacturing with greatest acceleration; Australia, Brazil, and France lead international growth
• Custom GPTs and Projects grow 19x, processing 20% of enterprise messages
• Growing gap: frontier workers send 6x more messages and leading companies use 7x more custom GPTs
• Sectors apply AI according to needs: technology in products, finance in compliance, retail in personalization
• Main challenge is organizational readiness: infrastructure, governance, leadership, and change management