We at OTE have been exploring the question of where AI can genuinely add value in the context of ERP systems for quite some time. And by “genuinely”, we really do mean genuinely. Not another chatbot that lets me ask questions about my products. Not a nice-looking summary of data that is already on the screen anyway. And not an AI feature that looks impressive but, three weeks later, no one quite remembers why we added it in the first place.
We are looking for use cases where we can say: Someone used to do this manually, it took a lot of time, and now AI takes over a significant part of the work. Apparently, sometimes all it takes is a little help from Brussels. Because while we are still looking for use cases like these, the EU has presented us with another bureaucratic monster in the form of the new Packaging and Packaging Waste Regulation, or PPWR.
And there it is: suddenly, we have a pretty good AI use case.
What Does a Packaging Regulation Have to Do with an ERP System?
Since 12 August 2026, large parts of the PPWR have applied directly across the EU. Further requirements will be introduced gradually over the coming years. Among other things, companies will need to take a closer look at packaging materials, recyclability, reuse and extended producer responsibility – or EPR. The specific obligations that apply to an individual company depend on its role, the packaging it uses, its supply chain and, in some cases, the country concerned. Each company therefore needs to assess which requirements apply to its own situation.
What interested us about the whole subject, however, was a different question: What does this mean for the data in the ERP system? And that is where things quickly become very concrete. Take a product that is sold in a cardboard box. There may also be wrapping paper, a plastic bag and perhaps a label. Business Central already knows exactly how many units of this product were sold, when they were sold, which customer bought them and which country they were delivered to. What Business Central often does not know is that the cardboard box weighs 280 grams, the paper 20 grams and the plastic bag 10 grams. With 50,000 units sold, that suddenly adds up to 14 tonnes of cardboard, one tonne of paper and 500 kilograms of plastic. If the products are sold in different countries, those quantities also need to be allocated accordingly.
And just like that, we are right in the middle of a classic ERP topic: master data plus transaction data equals reporting.
Microsoft Has Already Done Its Homework
When we took a closer look at the subject, we came across a small surprise: current versions of Dynamics 365 Business Central already include standard EPR functionality. EPR materials can be defined and the corresponding material compositions assigned to products. A product can therefore consist, for example, of 280 grams of cardboard, 20 grams of paper and 10 grams of plastic. Business Central then knows both this material composition and the actual inventory movements. This information can be used to derive the relevant EPR quantities and transactions for further reporting and analysis.
Microsoft Documentation on EPR in Business Central
Of course, this does not automatically mean that every PPWR requirement in every country is covered. Each company still needs to determine which obligations apply and which reports or declarations are required. But Microsoft has created something crucial: a sensible data structure. In theory, we could sit back at this point. If it were not for one small problem.
Now We Just Need to Maintain a Few Thousand Products
Setting up the EPR material composition in Business Central is not particularly difficult from a technical perspective. The question is: where does the data come from? With ten products, we can ask someone to find the relevant information and enter it. With 10,000 products, enthusiasm for that suggestion is likely to be somewhat limited. And that is exactly where the PPWR suddenly became an AI topic for us. In many cases, the information already exists. Suppliers and manufacturers provide product specifications. These arrive as PDFs, Excel files or in other formats and can contain extensive information about the respective product.
So the problem is often not that the data is missing. It is simply sitting unstructured in documents and somehow needs to make its way into the structured world of the ERP system. Until now, “somehow” has quite often meant that a person opens the document, finds the relevant information and types it into Business Central. And suddenly, we had our use case.
Dear AI, Could You Read the PDF, Please?
That is exactly what we are currently working on. We are developing an AI agent that reads existing product and supplier specifications, identifies the information they contain and then assigns it to the relevant products in Business Central in a structured way. From a PDF containing a detailed product specification, for example, the agent could identify that the packaging consists of 35 grams of cardboard and 12 grams of plastic. This information can then be assigned to the corresponding material composition in Business Central. And while we are at it, why stop at packaging?
Depending on the product and industry, these specifications contain a wide range of additional information that may also be relevant to product master data. The principle is always the same: a document contains structured and unstructured information that a person currently has to read, understand and then transfer into the ERP system. This is exactly the step an agent can support. And suddenly, we have an AI use case that meets three criteria that matter to us: The work already exists, it takes time, and no one will be sad if part of it is automated in future.
That is the kind of AI we like.
What About Older NAV or Business Central Versions?
There is, of course, one limitation. The new EPR functionality is only available in current versions of Business Central. Companies still using an older Business Central version or Dynamics NAV therefore first need to determine which requirements actually apply and how the necessary data can best be made available. As an interim solution, this could involve a separate data structure and corresponding reports, or it could form part of an already planned upgrade to a current version of Business Central.
From our perspective, the important thing is not for everyone to immediately start rebuilding their ERP system. What matters is asking early enough which data will be required in future and where that data is actually stored today. A new reporting obligation is inconvenient. A new reporting obligation followed by the realisation that you first need to track down the required data for 15,000 products is considerably more inconvenient.
Sometimes Bureaucracy Comes with Good Ideas
Of course, we would have explored AI agents and the automated transfer of product information even without another EU regulation. But the PPWR is a very good example of why we need solutions like these. On the one hand, requirements relating to product and master data are increasing. On the other hand, much of this information already exists somewhere within the company or with suppliers – just in the form of PDFs, Excel files or product specifications. Today, there is very often a person sitting in between, transferring information from A to B. And that is exactly where we believe AI should come in. Not because we desperately want to put an agent somewhere. But because we have found a process where we can actually use one.
So: thank you, EU.
We probably would have come up with this AI use case ourselves eventually. But apparently, a little bureaucratic monster provided just the extra motivation we needed.
