Business Central AI Example - Identifying and Understanding Inventory Shortages

For many years, I have been the one behind the keyboard when presenting demonstrations or providing user training within Microsoft Dynamics 365 Business Central. It has always been one of the most challenging and rewarding roles for me - allowing me to use my years of experience working within Business Central to help others run their businesses more effectively, operate more efficiently, and hopefully increase their appreciation and enjoyment from interacting with the system on a daily basis.

I feel like I was mostly successful :) However, amongst all of years of continuous change, enhancement, improvement, and technological evolution, there has always been one area where no level of advancement ever seemed to satisfy the typical user: the ease of interacting with the data in the system.

More specifically: How easy (or difficult) is it for an end user to generate the output required to answer a question they may have regarding the data that existed in the system. Questions around

• How to view a specific set of data?

• How to calculate a KPI or metric?

• How to build a new report?

• Or just how to simply obtain an answer to a specific question that came up during a meeting today?

And no matter how smoothly I was able to demonstrate obtaining the requested data, or how user-friendly the reporting tools were, it never seemed to be easy enough. And I get it - users are constantly trying to operate as efficiently as they can, and when they have to click 5 times or run through multiple steps to get the answer to a specific question, it can be frustrating. Especially when they are seeing it for the first time.

The day has finally arrived when in Business Central when users will no longer be able to say that it is too difficult to get the answers to their questions.

With the recent introduction of the AI tools that are now fully embedded within Business Central, the answers to questions are presented to the user simply by typing (or speaking) their questions into their favorite AI chat tool.

This is not a technical article, so I will not be explaining the specific technology or setup steps for enabling this functionality in each Business Central environment. But in a nutshell, this is accomplished by connecting an AI tool like ChatGPT or Claude to the Business Central environment using the MCP Server functionality that has been directly integrated into Business Central.

If you care to understand more about how this technology works, you can refer to this overview article on Microsoft Learn: Business Central MCP Server Overview and Setup - Business Central | Microsoft Learn. Be advised that Microsoft continues to make this setup easier and easier with each new release. For example, way back several months ago, this setup might have taken upwards of a full day to fully work out all of the kinks. Today, this setup can be completed in two hours or less.

My intent here is to provide a specific real-world example of how this technology can be used to interact with the data that resides in Business Central. This interaction includes not only how to obtain and present the data, but also to be able to better understand the data and what steps may be able to be taken to resolve an issue or improve a process.

While the ability to write data back to Business Central is possible, this example has been setup to only allow for the reading of data within Business Central. From a security and control perspective, this is key since we can use this technology without ever needing to worry about AI making incorrect or unauthorized changes to the data.

The remainder of this post will dive into a detailed real-world example of what this might look like for a typical end user. While I chose inventory shortage analysis as the example because it is something that most organizations have to deal with, it can be applied across any of the data that sits in Business Central.

Managing Inventory Shortages in Business Central using AI

A very common requirement for organizations that handle inventory is identifying and handling inventory shortages. Using standard ERP tools, this process would typically consist of running a report or viewing a planning page to identify the inventory shortages. Then, running additional reports or drilling-down to the underlying transactions. And finally, analyzing the entire data set to determine the best course of action and make decisions for how to resolve the shortages.

Let's take a look at what this process might look like using AI. These scenarios were tested in a demonstration company setup for Manufacturing, so the company utilizes Purchase Orders and Transfer Orders from other warehouses to re-supply raw materials and utilizes Production Orders to create the finished good items that will be shipped to the customer.

Note: Beyond manufacturing, this inventorty shortage scenario is also applicable to any company that manages inventory and uses any combination of the following within Business Central: Sales Orders, Service Orders, Assembly, Projects, or Manufacturing.

Let's start by identifying the existing shortages. We simply ask:

What Production Orders have shortages? What machine, assembly, or job is at risk?

Identifying where inventory shortages exist

Without prompting, the following information is also provided, which does its own analysis of the data in an attempt to identify the most urgent problem:

We could also take a slightly different approach and ask:

Which components are short? What exact Item is preventing production from moving forward?

Again, without additional prompting, it provides a more detailed analysis of the Items to provide insights into the most pressing issues or "hardest blockers":

After reviewing this information, perhaps we want to gather more granular information regarding the number of units needed:

How many units are needed, available, or short? What is the actual size of the shortage?:

Along with the above information, this question also provides further analysis regarding why the shortage exists for one of the key components:

We next work on resolving these issues. So we might want to identify if inventory may be available from someplace else:

Is there enough inventory somewhere else? Is the material unavailable at the production location but sitting at another location or bin?

Another relevant question would be to find out if inventory is already on the way:

Is replacement supply already coming? Is there an open Purchase Order, Planned Production Order, or other planned receipt?:

A "Bottom Line" analysis of each component shortage:

Next, we can work on the Purchase Orders, and with so many outstanding Purchase Orders, we need help to understand where we should start:

What document is supplying the material? Which Purchase Order or Production Order should the user investigate?

And a simplified list of where to begin is provided:

And of course, we need to know WHO should do this investigating:

Who is responsible for resolving it? Which buyer, planner, production employee, or department needs to take action?

Finally, and perhaps most importantly, we need to know which customers will be impacted by these inventory shortages:

What customer shipment is at risk? Which customer, machine, or promised shipment could be delayed?

This series of prompts and results is meant to provide examples of how you could interact with the data and the types of responses you can expect to receive.

Obviously, I could have created a much better prompt to avoid having to ask all of these questions separately. I could save this prompt, modify and enhance it as I gain more experience using it, and then apply it every morning just by telling it to use my standard prompt.

So what did we just see?

No need to run a series of reports! No need to open each Item page and drill-down to the underlying transactions! No need to understand how the Item and BOM structures are defined. Not even the need to open Business Central!

And not just the data presented, but also comprehensive analysis and suggestions provided on how to resolve the issues.

This is not just a new way to view data, it is a new way to analyze and manage the data. This is truly game-changing technology.

Wrap-up I have provided just one example to demonstrate how AI can be used within Business Central to simplify the process of generating, interpreting, understanding, and analyzing the data that sits in Business Central. This same functionality can be used to understand and analyze any data within Business Central - vendor performance, customer/item sales trends, project analysis, or financial performance.

As I noted earlier, it is also possible to utilize AI to take action on the data within Business Central through the use of AI Agents, but that is another topic for another day. For today, the key takeaway is that this functionality allows users to interact with their Business Central data using simple (and sometimes misspelled!) questions and prompts without the need for the traditional reporting and analysis tools we have all relied upon since we first started using ERP software!

Feel free to reach out if you are interested in setting this up and aren't sure how to get started.