The landscape of enterprise resource planning has shifted dramatically as we move through 2026, moving away from passive data collection toward proactive, executable intelligence. Supply chain professionals no longer have the patience for tools that merely point out a problem without offering a direct path to a solution. In this discussion, we explore the evolution of demand forecasting within the Microsoft Dynamics 365 Business Central ecosystem, focusing on how organizations are bridging the gap between sophisticated AI predictions and the daily grind of purchase orders and inventory management. Our guest is a veteran in ERP optimization who has spent years helping manufacturers and distributors turn “read-only” insights into “ready-to-act” results.
The conversation covers the inherent limitations of native ERP forecasting visuals and the necessity of interactive worksheets in modern planning. We delve into the mechanics of leveraging Azure AI without losing control over specific variables, such as algorithm selection and confidence levels. Furthermore, the discussion highlights the shift toward automating planning parameters like safety stock and reorder points, which allows teams to manage by exception rather than through manual, item-by-item updates.
Standard ERP tools often show sales forecasts as static, read-only charts that cannot be directly acted upon, leaving planners to manually bridge the gap. What are the primary frustrations you see when supply chain managers encounter these limitations in Business Central?
The most common frustration stems from the “glass wall” effect, where a manager can see a beautiful trend line generated by Azure AI but cannot touch the data to make it work for them. In the native Business Central Sales and Inventory Forecast, you might see a dip or a spike in a fact box, but there is no “apply” button to move those numbers into your actual planning logic. This forced disconnection means that even with sophisticated 2026-level AI running in the background, a human still has to manually transpose those insights into a separate spreadsheet or a different screen to update a reorder point. It feels incredibly inefficient to have a prediction engine that identifies a potential stockout but offers no path to prevent it without six extra clicks and manual data entry. Planners often feel they are babysitting the software rather than being empowered by it, leading to a lack of trust in the very AI meant to help them.
How does the Enhanced Forecasting Worksheet transform a simple visual into a functional tool that a planner can actually use to drive procurement?
This tool effectively takes the underlying Azure AI model that Business Central already uses and pulls it into an interactive environment where the data becomes “live.” Instead of looking at a static chart, the planner is presented with a worksheet where they can filter by item, item category, or even specific locations to get a granular view of the needs. The real power lies in the output options; you can push the resulting forecast directly into the standard Demand Forecast tables to feed your MRP and MPS runs. This eliminates the manual entry bottleneck entirely, turning a forecast from a suggestion into a hard requirement that the system can use to generate purchase orders. By providing a bridge between the prediction engine and the planning tables, it ensures that the forecast actually influences the supply chain rather than just sitting on a dashboard.
With six different forecasting algorithms available in this extension, how does an organization decide which approach is right for their specific product lines?
Choosing the right algorithm is about matching the mathematical model to the specific “personality” of an item’s sales history. For example, a fast-moving consumer good requires a different calculation than a slow-moving industrial component, and having the flexibility to choose between six distinct algorithms allows for that necessary precision. Planners can adjust the number of past periods and the confidence level to see how different models react to their historical data, essentially “tuning” the forecast until it matches reality. If the system flags an item because it doesn’t have enough posted sales history, the planner can immediately shift tactics rather than flying blind. This level of control ensures that a company isn’t trapped in a “one size fits all” prediction model that might over-forecast seasonal items or under-forecast steady sellers.
Could you walk us through the significance of using custom expressions to calculate planning parameters like safety stock and reorder points?
Moving away from static defaults is perhaps the biggest leap in efficiency for a purchasing manager in 2026. By building reusable expressions that reference variables like average daily usage and lead time, you can automate the maintenance of your Planning FastTab. Instead of a manager having to manually update the safety stock for 5,000 different SKUs, the worksheet calculates these values based on the latest forecast and pushes them directly onto the item cards. This means your reorder points are constantly evolving as sales patterns shift, which significantly reduces the risk of expensive overstock or devastating stockouts. It turns the item card from a stagnant record into a dynamic participant in the supply chain, reflecting the most current market conditions without constant human intervention.
For companies managing complex networks with multiple warehouses, why is it recommended to run these forecasts one location at a time rather than globally?
Handling forecasts location by location is a matter of maintaining data integrity and ensuring that local demand patterns aren’t “washed out” by a global average. According to the documentation for the Enhanced Forecasting Worksheet, running the forecast against a single location at a time ensures that the Azure AI model focuses specifically on the sales history of that site, providing a more accurate local reorder point. There is also a technical consideration: re-running an update against the same data adds to the prior result rather than replacing it, so a focused, one-location approach prevents messy data overlaps. This level of granularity is vital for organizations that might have high demand for a product in a metropolitan hub but very low movement in a regional satellite warehouse. It ensures that the right stock is in the right place, which is the ultimate goal of any distribution strategy.
Considering that this tool is available for free for a single user on the Microsoft Marketplace, how does that accessibility change the way small to mid-sized businesses approach AI?
The free single-user tier removes the traditional “pay-to-play” barrier that often kept smaller distributors from utilizing high-end AI forecasting. It allows a single planner to prove the ROI in a live Business Central online environment without an upfront financial commitment or a complex procurement process for new software. Because it builds on the native Azure AI that the company is already likely paying for as part of their Business Central license, it feels like a natural extension rather than a risky new investment. This allows a team to experiment with item categories or specific locations, seeing the tangible benefits of automated reorder points before deciding to scale the solution to more users. It’s a democratizing force in the ERP space, giving a ten-person distributor the same predictive power as a global enterprise.
What is your forecast for the future of AI-integrated demand planning within the ERP ecosystem over the next few years?
I anticipate that the line between “forecasting” and “execution” will continue to vanish until they are treated as a single, continuous process. We are moving toward a state where the ERP will not just suggest a reorder point but will automatically simulate the financial impact on cash flow before a human even looks at the worksheet. I expect we will see more integration between tools like the Enhanced Forecasting Worksheet and broader planning bundles, such as the Enhanced Planning Pack, which already includes tools for item planning reviews. The focus will shift from “how do we get a forecast?” to “how do we refine the expressions that govern our automated responses?” In short, the role of the supply chain planner will evolve from a data entry specialist into a “model manager” who oversees the logic that keeps the shelves full.