AI in ERP: what can it do and how to use it?

April 2025

ERP systems have long been an integral part of companies that value speed, efficiency and consistency in working with data. However, with the development of digital technologies, enterprise management systems are also evolving, now being supplemented by virtual assistants based on artificial intelligence. In a new article for speka.media, Anastasiia Katsan, a Senior consultant and systems architect of Microsoft Dynamics 365 Business Central at Innoware, explains how an AI assistant can be beneficial for ERP system users, as well as how it can automate routine business tasks, improve decision-making quality and significantly increase employee productivity.

Companies whose employees work faster and more efficiently are usually the ones that stay competitive. However, employees have recently gained a powerful assistant — artificial intelligence. It is now being implemented in many industries and business segments, and enterprise resource planning (ERP) automation systems are no exception. But such an assistant is not available to everyone yet. Let’s figure out how companies can use AI, for which tasks in ERP systems it is most effective, and who might fall behind this trend?

Imagine this situation: you come to work on Monday, and the first thing you hear from the management during the morning meeting is some disappointing news: «Our company missed the sales plan for the last quarter by 20%. If you want to stay on our team, you will have to work harder and more efficiently».

It’s easy to say, «work harder». You already often stay after work and put in extra hours. But there’s no way you’d be staying in the office every night… Just then, you receive an email from your IT director, reminding you that from now on, the ERP system your company uses has an intelligent virtual assistant powered by AI, which can take over not only routine tasks but even some creative work while you focus on other important tasks.

AI rushes to the rescue

The email from the IT director is encouraging, but it is still not very clear how to use this intelligent assistant. However, email does provide an answer to this question: it turns out that artificial intelligence can be used to train ERP system users, both new ones and long-time employees who want to master the latest features. All you need to do is ask: «how can I do this or that with your help…»? (Of course, AI will not completely replace the human instructor at all stages of system training, but it will still certainly make the process much easier). For example, if we ask in Copilot chat, which is available on any page of the cloud version of Microsoft Dynamics 365 Business Central, what sales reports can be generated, it will suggest several options and provide corresponding instructions for generating reports. Let’s assume that we have decided that this time we want to use the analysis mode.

• Generating a report

So, the first thing you need to do after your morning planning session is to generate a sales report for a specific period. Instead of clicking through menus and entering numerous parameters in the ERP system interface, you can simply give the AI the following task: «Sort all sales entries for December of last year, group them by product category, and include quantity, cost, and total sales amount». If for some reason you are not satisfied with the generated report, you can always modify it, for example, by giving a command like «group products by description instead of by code», or by manually changing the parameters, because the interface is user-friendly and similar to working with pivot tables in Excel. And if it turns out that the report is exactly what you need, you can export it to Excel, share it with colleagues, or save it for future use.

• Writing creative content

But just as you’ve sent the completed report to management, you received an email announcing that the company is launching sales of a large batch of entirely new products. You now need to write a technical description for each product, as well as a marketing text for the online sales site. The intelligent assistant can help here too, because you can give it the task of creating the appropriate text and even choose the tone: formal, casual, creative, etc. AI pulls data from the product attributes in the ERP system and quickly generates text in the required style based on that information.

• Processing a customer order

Just as you finish this task, the first customer orders for the new product start coming in. You don’t know the new nomenclature very well yet, but now you have an intelligent assistant: with its help you can quickly compile a list of products according to the customer’s order. For example, you can informally give the following command: «One stapler, one box of paper, and two flexible pencils with erasers». The system will generate an order with suggested product units such as a stationery stapler, office paper, and flexible pencils in the specified units of measure (if not specified, the configured one is used). If the selected items aren’t quite right, the list can be easily adjusted.

• Other features

In addition, AI can perform bank data reconciliation when we import a bank statement from a client-bank system and compare it to our actual transactions recorded in the ERP system. This enables to automate the reconciliation of transactions from the client-bank system with those recorded in the accounting system.

The example above isn’t science fiction – and it’s not even about the distant future! These technologies already exist today – this is what working in an ERP system looks like using the Copilot chatbot, which is already familiar to many from Windows 10/11 and the Microsoft 365 suite. Copilot has also been implemented both in the ERP system Microsoft Dynamics 365 Business Central and in the Microsoft Dynamics 365 ecosystem in general and is increasingly being integrated into various ERP functional modules that can be used in a variety of tasks.

However, Microsoft Copilot is not the only AI assistant in ERP. Artificial intelligence technologies are now being actively implemented in enterprise resource management systems by many well-known western providers. These systems are used by most Fortune 500 companies and are also familiar to many in Ukraine.

It should be noted that, like any other AI, Copilot does not think like a human, and sometimes it makes mistakes. This must be considered when working with it. So far, you cannot be 100% sure that the AI ​​has suggested the right path, so its decisions should be checked. However, when it comes to operations that we know the AI has performed accurately multiple times, it can be completely relied upon.

Overall, AI is constantly learning – from both internal corporate documents and publicly available sources. It is flexible in performing tasks and, moreover, has a good understanding of the Ukrainian language.

AI is not for everyone: who will not be able to adapt?

Thanks to the AI-powered virtual assistant, you were able to complete all your work on time, and the company was able to meet its sales plan in the next quarter. However, as mentioned earlier, these capabilities are not available to everyone. Artificial intelligence technologies are evolving rapidly and consistently, and it is quite likely that in six months or a year, tools like Copilot and similar AI assistants in ERP systems from other Western providers will make another significant leap in development. Such progress will benefit both users and developers: users will be able to process large volumes of work faster and more efficiently, while developers will gain tools to automate code writing.

At the same time, users of russian accounting systems like 1C/BAS and similar ones remain «out of trend». As a representative of one of the partners, who previously worked as a developer in this system, said: «For 1C users, I am artificial intelligence». Whatever they tell me to do, I do». In fact, 1C/BAS operates based on the same technologies that were used in Western systems 10-20 years ago. So far, no AI-driven innovations have been introduced.

Conclusions

Given the rapid pace of advancements in artificial intelligence, it’s safe to say that in just 3 to 5 years, any ERP system without AI will seem as outdated as keeping records on a wooden abacus does today. And since implementing an ERP system takes quite a long time – sometimes even years – it’s crucial to make the right choice from the start by considering for implementation only those solutions that are secure, modern and future-ready.

The original article is on speka.media