A week after Sam Altman sent out a “red code” memo due to competitive threats from Google, OpenAI published data indicating a sharp increase in the use of its AI tools over the past year. The company claims the results are excellent: the volume of ChatGPT messages has increased eightfold since November 2024, and employees in companies that have adopted the chatbot are saving up to an hour a day. OpenAI is being forced to rethink its position as a leader in corporate AI due to growing competitive pressure. Although, according to the Ramp AI index, nearly 36% of American companies are ChatGPT Enterprise customers, compared to 14.3% for Anthropic, the majority of OpenAI’s revenue still comes from consumer subscriptions—a base threatened by Google Gemini. Furthermore, OpenAI is increasingly facing competition from open AI model providers for corporate clients. Overall, OpenAI is counting on $1.4 trillion in investments for AI infrastructure development over the next few years, making the growth of corporate customers an essential part of its business model. “From an economic growth perspective, consumers are really important,” said OpenAI’s Chief Economist, Ronnie Chatterji. “But if you look at historically revolutionary technologies like the steam engine, it is the adoption and scaling of those technologies by companies that bring the greatest economic benefits.” The data provided by OpenAI shows that technology adoption among large enterprises is not only growing but also becoming more integrated into workflows. Organizations using the OpenAI API are consuming 320 times more “reasoning tokens” than a year ago, indirectly suggesting the use of AI for more complex tasks. However, this might only confirm active experimentation with new technology that doesn’t yield long-term benefits. Moreover, the growth in reasoning tokens correlates with increased energy consumption and can be costly for customers. OpenAI also described changes in how corporate clients use its tools. The use of AI assistants for workflow automation has grown 19-fold in the last year and now accounts for 20% of corporate messages. “This shows how capable people are in using this powerful technology and adapting it to their needs,” said OpenAI’s Chief Operating Officer, Brad Lightcap. According to OpenAI, AI integrations have led to significant time savings. Customers reported saving between 40 and 60 minutes daily thanks to OpenAI’s enterprise products, although this saving might not account for the time spent learning systems, writing prompts, or verifying AI outputs. OpenAI’s report notes that enterprise employees are increasingly using AI tools to augment their capabilities. The company claims that three-quarters of respondents say AI allows them to complete tasks, including technical ones, that they could not perform before. OpenAI reported a 36% increase in programming-related messages outside of engineering, IT, and research teams. While OpenAI promotes the idea that its technology empowers employees, it is worth noting that writing code with a chatbot can lead to an increase in security vulnerabilities and complications during software refinement. To detect errors, vulnerabilities, exploits, and other issues, the company released the Aardvark security research agent tool, which is currently in closed beta testing. Lightcap and Chatterji emphasized the “growing AI adoption gap,” with some “leading” employees using more tools than “laggards.” “There are companies that still view these systems as software, something I can buy, provide to my teams, and that’s where it ends,” said Lightcap. “And then there are companies that are truly starting to use it, almost like an operating system. It’s essentially a replatforming of many company operations.” OpenAI suggests that even the most active ChatGPT Enterprise users are not utilizing all the advanced tools available to them, such as data mining, reasoning analysis, or search. Lightcap believes that the full implementation of AI systems requires a significant shift in employee mindset and deeper integration of AI with corporate data and processes. It will take time to better understand the capabilities and restructure workflows.
About the Author
