# Microsoft Transforms its Cloud Supply Chain with Optimization and Generative AI

## Метаданные

- **Канал:** Microsoft Research
- **YouTube:** https://www.youtube.com/watch?v=WeLQEMGx9_0
- **Дата:** 16.04.2026
- **Длительность:** 2:05
- **Просмотры:** 460
- **Источник:** https://ekstraktznaniy.ru/video/46442

## Описание

Microsoft transformed its cloud supply chain with the Intelligent Fulfillment Service (IFS), a breakthrough platform that combines machine learning, mathematical optimization, and generative AI. By optimizing global shipment planning, IFS cuts cycle times by more than half and delivers substantial financial benefits. Its large-language-model-powered assistant, built on the pioneering OptiGuide framework, brings real-time explainability and scenario exploration to planners, significantly reducing the fulfillment team's workload by compressing decision cycles from days to minutes.

See more: https://www.microsoft.com/en-us/resea...

## Транскрипт

### Segment 1 (00:00 - 02:00) []

Our online world is powered by the cloud. At Microsoft, we orchestrate one of the largest cloud supply chains on the planet. Every day, we determine when cloud capacity comes online, where it's deployed, and which hardware to ship. Given what's happening in AI today, we are engaged in a massive infrastructure buildout. — In fact, it might be the largest engineering effort ever undertaken — by mankind. And with that comes a whole host of optimization problems that are at a scale also never encountered before. For years, cloud fulfillment relied on manual spreadsheet-driven planning. But as demand surged, plan generation became slower and harder to validate, and therefore fragile in the face of disruption. Microsoft Research and the Cloud Supply Chain organization created the Intelligent Fulfillment Service to balance policy tradeoffs and deliver capacity faster. IFS makes millions of simultaneous decisions — using mathematical optimization with decomposition techniques. While globally optimal, the resulting plans often appeared counterintuitive to supply chain planners when they viewed them at the individual demand level. It was hard for the planners to understand why the system recommended a plan. I think with Gen AI, we saw an opportunity to make this complex reasoning be easily accessible using natural language question. We augmented IFS with a generative AI layer called OptiGuide. So planners can understand decisions and explore what-if scenarios without needing to know the math. Together, optimization and generative AI transform cloud supply chain at Microsoft. From manual planning to automated decisions that are fast, explainable, and trusted. Importantly, this work doesn't replace operations research, but extends it, thus making it easier for people who are less familiar with new methods to integrate and adopt it in real-world settings.