For decades, industrial and equipment manufacturers have relied on historical demand as the foundation for forecasting. Looking at past sales patterns and projecting them forward was often enough to guide purchasing, production planning and inventory management.
But in today’s environment, that model breaks down quickly. The aftershocks of the COVID pandemic were followed by supply chain disruptions, sudden spikes and drops in demand, political and global instability, and ongoing uncertainty around tariffs. In this kind of landscape, relying purely on historical trends can lead to costly miscalculations.
Adapting can feel daunting, and it may even raise the question of whether accurate forecasting is still possible. The good news is that new AI tools have emerged to help manufacturers navigate demand planning uncertainty.
As a result, the conversation has shifted from “How do we forecast?” to “How do we forecast intelligently?”
Why AI alone isn’t enough
Artificial intelligence is increasingly part of that answer, but it isn’t a complete solution on its own. While AI excels at analyzing large volumes of data and identifying patterns, it doesn’t inherently understand the business context behind those patterns.
That distinction matters. A forecasting model might detect a sudden increase or decrease in demand, but it may not understand why that shift occurred. For example, a temporary drop in demand caused by tariffs, or a short-term spike created by customers accelerating orders ahead of anticipated price increases, can distort the data used to predict demand.
Traditional ERP forecasting tools typically rely on historical demand data within the system. They apply smoothing or averaging methods and then generate planned orders based on established parameters. When demand signals are distorted, however, those models can produce misleading results.
Without human judgment to interpret the broader context, the result may be unnecessary inventory investment and excess product once demand returns to normal. In a volatile market, understanding long-term patterns rather than reacting to short-term turbulence becomes critical.
Where AI adds value in demand planning for manufacturer
Although AI cannot replace human insight, it can significantly improve the forecasting process by helping planners analyze data more effectively.
AI is particularly strong at identifying patterns, highlighting deviations from expected trends and surfacing anomalies within large datasets. Instead of manually sorting through spreadsheets or attempting to smooth complex data in Excel, planners can use AI to detect unusual spikes, isolate irregular growth patterns and flag areas that warrant closer examination.
Once these insights are surfaced, human planners can apply their experience and judgment to determine whether the changes represent lasting market shifts or temporary disruptions.
In other words, AI accelerates the discovery of meaningful signals while people provide the interpretation to turn them into actionable decisions.
Expanding forecast intelligence with Microsoft’s Power Platform
Historically, demand forecasting functionality has lived inside the four walls of Microsoft Dynamics 365 Finance & Supply Chain Management. While useful, this approach limits forecasting to the data stored within the ERP environment. Historical data archived in data warehouses or information from external systems often remain disconnected from the forecasting process.
Forecasting capabilities are expanding through Microsoft’s Power Platform, which can be used to:
- Create a Power App to enter forecast adjustments
- Use Power BI to visualize demand trends
- Use Azure Machine Learning for prediction models
- Use Power Automate to trigger alerts when demand spikes
These tools make it easier for planners to collaborate with AI rather than simply accept its recommendations.
Consider a situation where demand for a particular item has grown steadily over the past decade, only to experience a sudden short-term spike. AI can quickly identify that divergence from the long-term pattern. A planner can then evaluate whether the spike represents a true shift in market demand or simply a temporary anomaly that should be excluded from the forecast. This allows organizations to benefit from both computational power and business judgment.
Available: demand planning application for manufacturers
The prebuilt Demand Planning application built on Power Apps runs on Power Platform and integrates with Dynamics 365 Finance & Supply Chain Management to support more flexible demand and supply planning. The functionality sits inside one application and reduces the customization required if you were to build separately on Power Platform. Manufacturers can use the application to build Power BI dashboards, add custom features and integrate external data sources.
The Demand Planning application provides a no-code environment that allows planners to build demand models, run “what-if” scenarios and collaborate on forecasts across teams. AI-driven forecasting capabilities can automatically tune parameters for improved predictions, while external signals such as promotions, stockouts and market events can be incorporated to improve accuracy.
The application also enables planners to adjust forecasts at higher levels (such as product families or regions) and then drill down to individual SKUs as needed.
While these capabilities are powerful, the technology is evolving. Documentation and implementation maturity continue to develop, and some still manufacturers may choose to evaluate additional third-party tools depending on their planning requirements.
Built-in D365 planning tools still play an important role
Even without advanced Power Platform capabilities, Dynamics 365 Finance & Supply Chain Management includes built-in planning tools that help manufacturers manage demand volatility. However, these native capabilities still rely on thoughtful configuration and human oversight.
For example, safety stock calculations, planning horizons and parameter-driven logic all influence the accuracy of system-generated forecasts.
Imagine a procurement agent reviewing a purchase order scheduled to arrive in 32 days while running a forecast that covers a 30-day window. A purely automated system might interpret the absence of orders within that 30-day range as a signal to generate a new purchase order.
By configuring planning parameters to include a buffer period before and after the forecasting window, planners can ensure the system captures activity that would otherwise fall outside the model’s view.
This type of intelligent configuration highlights an important reality: even the most advanced tools perform best when guided by experienced planners who understand the operational context behind the data.
Using forecasting to build a more resilient business
By integrating decision-making processes with platforms like Dynamics 365 Finance & Supply Chain Management and Power Platform, manufacturers can better align forecasting with financial planning, procurement and production strategies.
Used effectively, AI becomes a powerful support system for demand planning. It acts as a pattern recognition engine, an anomaly detection tool and a data accelerator that allows teams to work faster and more confidently. Manufacturers that embrace this can respond to volatility without overreacting to short-term disruptions, helping them make more resilient planning decisions.
Of course, successfully implementing these tools often requires experienced guidance. Sikich can help you identify the right solutions, avoid common pitfalls and maximize the value of both AI and human insight.
This publication contains general information only and Sikich is not, by means of this publication, rendering accounting, business, financial, investment, legal, tax, or any other professional advice or services. This publication is not a substitute for such professional advice or services, nor should you use it as a basis for any decision, action or omission that may affect you or your business. Before making any decision, taking any action or omitting an action that may affect you or your business, you should consult a qualified professional advisor. In addition, this publication may contain certain content generated by an artificial intelligence (AI) language model. You acknowledge that Sikich shall not be responsible for any loss sustained by you or any person who relies on this publication.
