Microsoft’s new decision-scoring model, Microsoft-Decision-1, is already being experimented with inside both Bing Search and Copilot, days after the company announced it. The model is built for structured decision work — routing, classification, prioritisation, verification and workflow control — and Microsoft says it beats both large language models and rival decision models on latency and quality for those tasks.
The experiment surfaced publicly when members of the Bing team acknowledged it on social media this week: one engineer posted that the team is experimenting with the model in Copilot and is “pretty good” impressed with how quickly it improved, while a Bing spokesperson replied to questions with the classic “we are always experimenting.”
Why should search watchers care about an enterprise-sounding model? Because decision-scoring is exactly the machinery an AI search engine runs constantly — deciding which sources to trust, which answer path to take, and which results to verify before responding. If Decision-1 earns a permanent place in that pipeline, expect quieter, faster changes in how Bing and Copilot choose what to show, long before any of it gets a launch post.
Source: Search Engine Roundtable
0 Comments