Microsoft is changing its approach to its flagship AI product, Microsoft 365 Copilot, by using a wider range of AI models. Launched in March 2023 with OpenAI's GPT-4 at its core, the company is now adding its own models like Phi-4 and other third-party models. This shift is aimed at reducing reliance on OpenAI, cutting costs and improving performance, especially for enterprise users who are concerned about speed and affordability.
The main goal is to make the AI assistant more cost-effective and user-friendly for businesses. Microsoft is working on smaller in-house models and modifying open-weight models to strike a balance between performance and cost. While exploring new options, Microsoft continues to partner with OpenAI for advanced models, ensuring access to the latest AI technologies.
The decision to expand AI models comes as Microsoft faces challenges with the pricing and utility of 365 Copilot. A Gartner survey found that many IT companies are still in the trial phase of using the tool. Despite these issues, the adoption rate is promising, with experts predicting over 10 million paid users by the end of the year. Microsoft also shared that 70% of Fortune 500 companies are already using 365 Copilot. By diversifying its AI models, Microsoft aims to offer better value and long-term success in the competitive AI market.
