Microsoft's new feature makes AI agents 4 times faster and less expensive

By intelligently organizing data, Microsoft increases the speed of artificial intelligence agents by 4 times and reduces token consumption by 30%. The post Microsoft's new feature makes artificial intelligence agents 4 times faster and less expensive appeared first on Safir Arjan.

By intelligently organizing data, Microsoft increases the speed of artificial intelligence agents by 4 times and reduces token consumption by 30%.

Become a special member of Arjan Ambassador in order to benefit from and access special facilities and different departments in Arjan Ambassador.

Microsoft recently developed a new experimental capability to categorize data and change the way it is presented to agents, which can make AI tools up to 4x faster and optimize token consumption by up to 30%.

The continuous and rapid improvement of massive language models in each update provides the basis for better and smoother agent performance. Of course, more advanced models do not only lead to better agents; Increasing efficiency is also one of the benefits of such updates.

In the latest effort in this direction, Microsoft announced that only by changing the structure of information organization and the way they are presented to artificial intelligence agents, it was able to significantly improve the performance of these tools; The result obtained without extensive changes in the agent controller model.

In the latest announcement related to Copilot and artificial intelligence agents, Microsoft announced that it was able to increase their response speed by four times; While the amount of token consumption has also decreased by 30% compared to the past.

This new feature is developed for the Microsoft SharePoint platform using the Work IQ engine; An engine that can organize data in a way that agents can read it faster and more efficiently.

For example, information about a project may be scattered across hundreds of emails, team members' conversations on different platforms such as Teams, memos, technical documents, and other sources. Normally, an AI agent must collect and process data from various sources in order to answer user queries or perform more complex tasks.

Using Microsoft's new feature, this data is organized and grouped into files that agents can read more quickly. In the future, users will be able to access these files and refer to their content. Also, the ability to automatically update data files is considered so that agents can collect and add new information from various sources.

Microsoft has shown in an article that using this method to categorize organizational data can improve agents' access to information, response accuracy and token consumption efficiency. The main strength of such a system is answering the questions that the information needed to answer them is scattered in different sources and no single source is sufficient to provide the answer.

This feature will initially be available to institutions and organizations that have registered in Microsoft's Frontier test program. Also, it is possible to add it to SharePoint based sites.

In the near future, Microsoft plans to add this feature to Copilot Studio agents and then expand its use to more parts of the Microsoft 365 platform. However, an official and exact schedule for the release of this feature in the next stages has not been announced yet.

Amir is very interested in common fields of biology and technology and is active in the field of writing educational content and reviewing services and applications.