Case Study
Sight Machine: The Factory Schedule That Learned to Replan Before the Next Meeting
AI-generated optimization models turned real-time disruptions into updated production schedules.
The best schedule is not the one with the prettiest baseline; it is the one that can absorb live constraints and produce a feasible next move quickly.
Frequent disruption forced manual replanning meetings 10–15 times per week and consumed operator and planner expertise.
A factory schedule can become obsolete before the ink metaphorically dries. The beverage manufacturer in this Microsoft story was replanning 10 to 15 times each week. Sight Machine's OptiMind combined current orders, plant data, constraints, and optimization so the schedule could respond to disruption without gathering everyone for another emergency spreadsheet summit.
OptiMind used Microsoft Foundry, natural-language inputs, real-time plant data, and generated optimization models to refresh production schedules when conditions changed.
Microsoft reports the customer outcomes and the page's detailed result section is used here. For project managers, this is dynamic scheduling: maintain trusted constraint data, define the objective function, test feasibility, and keep operators able to challenge the recommendation. Optimization is powerful, but a mathematically elegant plan that ignores a maintenance window is still just expensive fiction.