Utilizing ERP Data

ERP systems are effective at detailing production plans. They are far less effective at detailing how work actually moves through a manufacturing environment once variability, backlog, and product complexity are introduced. A growing backlog indicated that our ERP system was no longer providing a reliable view of where work truly sat in the production process. Leadership decisions were increasingly driven by assumptions, manual reconciliation, and spot checks rather than accurate operational information. 

To regain clarity, I analyzed our ERP database at the SQL level and linked work orders and products to their last completed operation and their intended next step—capturing who worked on it, when, and for how long. The resulting data was structured in Excel to provide the expected location of every part on the shop floor. By layering in required ship dates, associated revenue, and part types, management was given a summary-level view of backlog growth, bottlenecks, late orders, and revenue exposure. 

Once a clearer baseline was established, I built a deeper operational view by walking the shop floor, speaking directly with personnel, and verifying the physical status of products. While accurate, this approach was not maintainable at scale. To address this, I leveraged a secured enterprise AI environment to codify that operational logic and apply it consistently to the ERP-derived data—automatically highlighting meaningful changes and exceptions. This reduced the number of products requiring daily follow-up from hundreds to a manageable handful. 

Here the data is summarized by product type and work station to reflect how work actually flows through our manufacturing process. While this breakdown would not be appropriate for every environment, it aligns closely with how constraints, handoffs, and backlog form on our shop floor—allowing issues to be identified and addressed where they truly occur. 

On a separate view, individual work orders are grouped by their most likely next operation based on where they were last recorded. For example, work last seen at Injection is expected to progress to Cooling next. Organizing work this way shifts the focus from static system states to anticipated flow, making it easier to spot stalled work, emerging constraints, and misaligned priorities before they impact delivery.