Microsoft to reduce OpenAI for its 365 Copilot, here’s why

Microsoft is actively pursuing strategies to lessen its reliance on OpenAI for its premier artificial intelligence product, Microsoft 365 Copilot.

This initiative is part of a larger strategy aimed at reducing expenses and enhancing performance for enterprise clients.

Although Microsoft has maintained a long-standing partnership with OpenAI, which included the integration of its GPT-4 model into 365 Copilot upon its launch in March 2023, the company is now investigating alternative AI models.

According to sources, Microsoft is working to integrate both proprietary and third-party AI models into 365 Copilot. A primary concern for Microsoft has been the costs and processing speed associated with OpenAI’s models, which could impact the experience of enterprise users.

Read more: Microsoft introduces new file sharing feature between iPhones, Windows PCs

A spokesperson for Microsoft highlighted that OpenAI continues to be a significant partner, particularly for the most sophisticated AI models. “We utilize a variety of models from both OpenAI and Microsoft based on the specific product and user experience,” the spokesperson stated.

To support this objective, Microsoft has been developing its own smaller AI models, such as Phi-4. Furthermore, the company is focused on customizing open-weight models to enhance the performance of 365 Copilot. By integrating these new models, Microsoft aims to optimize AI processes and lower operational costs, potentially resulting in savings for customers.

This strategic shift to decrease reliance on OpenAI is part of a broader effort to manage expenses in a highly competitive AI market.

In the long term, this strategy could render 365 Copilot more cost-effective and efficient, ultimately benefiting both Microsoft and its clientele.