Here’s why Microsoft is reducing OpenAI for its 365 Copilot.

Microsoft 365 Copilot is its flagship artificial intelligence product, and the company is aggressively looking for ways to reduce its dependency on OpenAI.

This project is a component of a broader plan to lower costs and improve corporate client performance.

Despite its long-standing collaboration with OpenAI, which included integrating its GPT-4 model into 365 Copilot when it was first released in March 2023, Microsoft is currently looking into other AI models.

Microsoft is reportedly working on integrating third-party and proprietary AI models with 365 Copilot. Microsoft’s main worries have been the expenses and processing speed of OpenAI’s models, which may affect enterprise users’ experiences.

OpenAI is an important partner, especially for the most advanced AI models, a Microsoft representative emphasized. According to the spokesman, “We use a variety of models from both OpenAI and Microsoft depending on the specific product and user experience.”

Microsoft has been working to assist this goal by creating its own tiny AI models, such as Phi-4. Additionally, the business is working to improve 365 Copilot’s performance by adapting open-weight models. Microsoft hopes to reduce operating expenses and optimize AI operations by incorporating these new models, which could save clients money.

A larger attempt to control costs in a fiercely competitive AI sector includes this tactical change to lessen dependency on OpenAI.

Over time, this approach might make 365 Copilot more economical and successful, which would eventually help Microsoft and its customers.