Announcing Microsoft Web IQ (blogs.bing.com)

Microsoft has launched Web IQ, a suite of AI-native grounding APIs designed to enhance the capabilities of AI systems by providing access to fresh, high-quality web data. Transitioning from traditional AI frameworks, which often rely solely on model capabilities, Web IQ emphasizes the importance of dynamic information retrieval and processing. This system is structured to meet the challenging demands of multi-step reasoning, enabling AI agents to efficiently access and utilize real-time data from a vast range of online sources, including web pages, images, and videos. The significance of Web IQ lies in its re-architected design, which combines years of expertise from Microsoft’s Bing with a focus on agentic operations, incorporating distributed systems strategies to optimize data retrieval. It utilizes an advanced embedding model for content analysis and ranking, ensuring high-quality results while operating on a passage-level rather than document-level basis to maximize information density. Web IQ aims to achieve superior latency and grounding satisfaction metrics, making it a promising tool for powering AI applications that need to interact with ever-changing real-world information efficiently and effectively.