Prometheus operates by using Bing and GPT to produce a sequence of internal queries through Bing Orchestrator. The model then delivers an accurate and in-depth response to the user's inquiry while taking into account the context of the conversation, all within milliseconds.
Selecting the relevant internal queries and utilizing the corresponding Bing search results is a crucial component of Prometheus. This process allows the model to access up-to-date information and reduce inaccuracies. In other words, the model grounds itself in the data provided by Bing by using its reasoning skills, aided by the Bing Orchestrator. A diagram is included in the post to provide an overview of how Prometheus functions.
Jordi Ribas, Corporate Vice President of Search & AI at Microsoft, stated in the blog post that his team faced challenges while developing the ideal Bing search product, and they ultimately designed a new Bing UX that incorporates both Search and Chat in a single interface to provide a unified experience for users. This was because certain queries would perform better in the classical or Search mode, while others would do better in a conversational or Chat mode.