OpenAI's 2025 State of Enterprise AI report reveals that artificial intelligence is increasingly embedded in enterprise workflows across industries and regions. The findings draw on de-identified and aggregated usage data from OpenAI's enterprise customers and a survey of 9,000 workers across almost 100 enterprises. No OpenAI employee reviewed individual enterprise customer data.
Weekly ChatGPT Enterprise messages have grown approximately 8 times since November 2024, while the average worker sends 30% more messages. Usage of structured workflows such as Custom GPTs and Projects expanded approximately 19 times year-to-date, highlighting a shift toward repeatable, integrated processes. Average reasoning token consumption per organization increased about 320 times over 12 months, reflecting broader deployment of advanced models in internal systems and products.
Workers reported measurable operational benefits. Seventy-five percent of surveyed workers report that using AI at work has improved either the speed or quality of their output. On average, workers saved 40–60 minutes per workday, with data science, engineering, and communications workers saving 60–80 minutes daily. The heaviest users reported saving more than 10 hours per week. Across departments, 87% of IT workers reported faster issue resolution, 85% of marketing and product teams reported quicker campaign execution, 75% of HR professionals noted improved employee engagement, and 73% of engineers reported faster code delivery.
AI is also enabling employees to perform tasks outside their traditional roles. Seventy-five percent of workers report being able to complete tasks they previously could not perform, including programming support and code review, spreadsheet analysis and automation, technical tool development and troubleshooting, and custom GPT or agent design. Coding-related activity outside engineering, IT, and research increased by 36% in the six months preceding the report.
Industries showing the fastest growth include technology (11 times year-over-year), healthcare (8 times), and manufacturing (7 times), with the median sector expanding more than 6 times. Growth in paying business customers was particularly strong in Australia (187%), Brazil (161%), the Netherlands (153%), and France (146%).
On the other hand, Anthropic's independent analysis of 100,000 Claude conversations found that Claude estimated an 80% reduction in task completion time. Anthropic cautions, however, that these estimates likely overstate productivity effects because they do not account for time spent refining outputs, validating quality, or working across multiple sessions. Validation against software development tasks showed moderate correlation between Claude's estimates and actual tracked completion times.
The report documents that enterprise AI adoption is expanding across both the number of workers using AI and the depth of integration into workflows, with workers engaging with advanced tools and structured workflows reporting larger productivity gains.
Read next: EU Accepts Meta’s Updated Ad Model for January 2026 Rollout
Weekly ChatGPT Enterprise messages have grown approximately 8 times since November 2024, while the average worker sends 30% more messages. Usage of structured workflows such as Custom GPTs and Projects expanded approximately 19 times year-to-date, highlighting a shift toward repeatable, integrated processes. Average reasoning token consumption per organization increased about 320 times over 12 months, reflecting broader deployment of advanced models in internal systems and products.
 |
| Productivity gains increase with intensity of AI use |
Workers reported measurable operational benefits. Seventy-five percent of surveyed workers report that using AI at work has improved either the speed or quality of their output. On average, workers saved 40–60 minutes per workday, with data science, engineering, and communications workers saving 60–80 minutes daily. The heaviest users reported saving more than 10 hours per week. Across departments, 87% of IT workers reported faster issue resolution, 85% of marketing and product teams reported quicker campaign execution, 75% of HR professionals noted improved employee engagement, and 73% of engineers reported faster code delivery.
AI is also enabling employees to perform tasks outside their traditional roles. Seventy-five percent of workers report being able to complete tasks they previously could not perform, including programming support and code review, spreadsheet analysis and automation, technical tool development and troubleshooting, and custom GPT or agent design. Coding-related activity outside engineering, IT, and research increased by 36% in the six months preceding the report.
Industries showing the fastest growth include technology (11 times year-over-year), healthcare (8 times), and manufacturing (7 times), with the median sector expanding more than 6 times. Growth in paying business customers was particularly strong in Australia (187%), Brazil (161%), the Netherlands (153%), and France (146%).
On the other hand, Anthropic's independent analysis of 100,000 Claude conversations found that Claude estimated an 80% reduction in task completion time. Anthropic cautions, however, that these estimates likely overstate productivity effects because they do not account for time spent refining outputs, validating quality, or working across multiple sessions. Validation against software development tasks showed moderate correlation between Claude's estimates and actual tracked completion times.
The report documents that enterprise AI adoption is expanding across both the number of workers using AI and the depth of integration into workflows, with workers engaging with advanced tools and structured workflows reporting larger productivity gains.
Read next: EU Accepts Meta’s Updated Ad Model for January 2026 Rollout
