Enterprise use of artificial intelligence is expanding at a rapid pace as organisations move beyond experimentation and begin embedding AI into core systems and daily operations. According to OpenAI’s 2025 State of Enterprise AI report, more than one million business customers now use its tools, with activity rising across functions, industries, and regions. The report notes that ChatGPT workplace seats have surpassed seven million, while ChatGPT Enterprise seats have grown nine times year-on-year. Weekly message volume from enterprise users has also increased eight times since November 2024, showing both wider adoption and intensified usage.
The report highlights that AI is transitioning from a consumer-facing novelty to a foundational technology inside large organisations. Many firms are now applying AI to multi-step workflows that require precision, security, and scale. This shift mirrors the adoption paths of earlier general-purpose technologies, where the greatest value emerged only after companies built repeatable, production-grade applications.
Workflow tools and API usage are reshaping enterprise productivity
One of the clearest trends in the report is the rise of Custom GPTs and Projects. These tools allow teams to build reusable assistants tailored to internal processes, institutional knowledge, or system integrations. Weekly users of Custom GPTs and Projects have increased nineteen times this year. Around 20 per cent of all enterprise messages now flow through these customised interfaces. OpenAI notes that some organisations, such as BBVA, run thousands of internal GPTs to support repeatable tasks at scale.
Developer adoption is also accelerating. More than 9,000 organisations have processed over 10 billion tokens through the API, and nearly 200 firms have surpassed a trillion tokens. Average reasoning token consumption per organisation is up 320 times in the past year, indicating that companies are integrating more capable models into production systems. Codex usage has grown quickly too, with significant increases in automated code generation, testing, and debugging.
These developments translate into measurable productivity gains. Survey data from nearly 100 enterprises shows that 75 per cent of workers report improvements in the speed or quality of output when using AI. On average, users save between 40 and 60 minutes per active day. Workers in data science, engineering, and communications roles report even higher savings, up to 80 minutes a day. Functions such as IT, marketing, HR, and engineering also record substantial improvements, from faster issue resolution to shorter production cycles.
The report emphasises that AI is enabling workers to perform tasks that previously required specialist skills. Around 75 per cent of surveyed employees say they can now complete new technical work, including coding, spreadsheet automation, and building simple internal tools. Coding-related messages have increased across all departments, not only technical functions.
Adoption accelerates worldwide but a usage divide is emerging
AI uptake is broad-based across industries, with the median sector growing six times year-on-year. Technology, healthcare, and manufacturing show the fastest growth, while finance and professional services represent the largest current user base. API usage is diversifying as well, with non-technology firms increasing adoption five times over the past year. Customer support, content generation, and workflow automation are becoming common AI-driven functions across sectors.
Geographic adoption is widening too. Countries such as Australia, Brazil, the Netherlands, and France recorded growth rates above 140 per cent in the number of paying business customers. The United Kingdom and Germany now rank among the largest ChatGPT Enterprise markets outside the United States, while Japan leads international API usage.
Despite this momentum, the report highlights a widening gap between leading and lagging users. Frontier workers, defined as those in the top five per cent of usage intensity, send six times more messages than the median user and generate far more activity in advanced tools such as data analysis and reasoning features. At the organisational level, frontier firms produce twice as many messages per seat and send seven times more messages to GPTs. This reflects deeper workflow integration, greater standardisation, and stronger internal support structures.
The report warns that many enterprise users still have not touched core capabilities: 19 per cent of monthly active users have never used data analysis tools, 14 per cent have not used reasoning, and 12 per cent have not used search. This suggests significant headroom for growth as organisations refine training, governance, and change management frameworks.
Case studies show financial impact and new operating models
The report also presents case studies demonstrating how AI is delivering tangible business results across different sectors. Intercom used OpenAI’s Realtime API to build Fin Voice, a low-latency voice agent for customer support, reducing latency by 48 per cent and resolving more than half of calls without human intervention. Lowe’s deployed Mylow, an AI assistant for both online customers and in-store associates, answering nearly one million queries per month and more than doubling online conversion rates.
Indeed introduced AI-driven matching and personalised explanations for job seekers, increasing application starts by 20 per cent and improving hiring outcomes. BBVA automated legal queries related to corporate signatory checks, reducing manual workloads and achieving material cost savings. Oscar Health rolled out AI assistants integrated with medical records and claims systems to handle benefits and healthcare queries in real time. Moderna used AI to compress a core analytical step in its Target Product Profile development process from weeks to hours.
Across these examples, AI is linked to improved customer experience, lower operational costs, increased employee capacity, and faster product development. External research cited in the report notes that AI leaders achieve stronger financial performance, including higher revenue growth, total shareholder return, and profit margins.
The report concludes that enterprise AI remains in its early stages, with significant opportunities ahead as organisations shift from simple use cases to deeper workflow integration and market-facing innovation.
