Why Many Companies Aren’t Ready for AI Yet – and Why That’s Perfectly Fine
In my first article, I argued that technology should not be the starting point – the objective should. Only once it is clear what needs to be improved can you determine whether artificial intelligence can actually provide meaningful support. Let’s assume that step has already been taken: there is a specific use case. The company has decided that AI could, in principle, be the right solution. So surely it is time to get started. Right? This is exactly where I often encounter the same reaction in projects: “We’re probably not ready for that yet.”
Interestingly, this statement rarely has much to do with the software being used or the size of the company. I hear it just as often from companies that have been working with a modern ERP system such as Microsoft Dynamics 365 Business Central for years as I do from companies that are only beginning to digitise their processes. The real uncertainty usually comes down to a different question:
Does AI actually know enough about our company to provide meaningful support?
From my perspective, that is precisely the crucial point.
AI Doesn’t Know Your Company
Modern AI systems possess an impressive amount of knowledge. They can write texts, summarise documents, identify patterns and explain complex relationships. They understand how ERP systems generally work, are familiar with typical business processes and can interpret a wide range of specialist terminology. What they do not know, however, is your company:
- AI does not know which customers are particularly important to your business.
- It does not know your suppliers.
- It does not know why certain processes have evolved over the years or why some of them deliberately work differently from what a textbook might recommend.
Microsoft Dynamics 365 Business Central does not automatically understand these relationships either. The ERP system manages information. But the meaning of that information only emerges from the context of your business. That is precisely why AI cannot simply “get started”. First, it needs access to the knowledge your company uses every day.
There Is a Big Difference Between Data and Knowledge
When people talk about AI, sooner or later the word “data” almost always comes up. This often creates the impression that good data alone is enough for AI to produce meaningful results. In my experience, that view falls short. There is a crucial difference between data and knowledge. Take a product in food wholesale. Business Central knows the product number, prices, inventory levels and previous sales. Initially, AI sees exactly the same information.
What it doesn’t know:
- Why does this product sell significantly more every summer?
- Why is it stocked several weeks in advance?
- Why do we order from this supplier despite higher purchase prices?
- Why is a particular customer treated differently from other customers?
For employees, these relationships are often obvious. For AI, they initially do not exist. It needs context. Only when data is supplemented by rules, processes and experience does it become knowledge that AI can actually use.
A reliable data foundation remains essential. In our article on common digitalisation mistakes, we look at why digital transformation projects often fail not because of the technology, but because of data quality, disconnected processes and structures that have evolved over time.
Common Digitalisation Mistakes in Food Wholesale, Retail and Distribution
Read Bok-Him Lin’s article.
Why Apparently Simple Questions Suddenly Become Difficult
This becomes particularly clear with questions that seem perfectly straightforward in everyday business.
A managing director might ask, for example: “Which customers order irregularly?”
To the sales team, the question may seem perfectly clear. For AI, it is surprisingly complex:
- What does “irregularly” mean?
- Is the customer ordering less frequently than before?
- Has their ordering pattern changed?
- Are seasonal fluctuations relevant?
- Or are we trying to identify customers whose purchasing behaviour is currently changing?
Any of these interpretations could be correct. It depends on what the question means within your company. The same applies to questions such as:
- Which suppliers are particularly reliable?
- Which products are performing well?
- Which customers are particularly important to us?
- Which orders should be prioritised?
For employees, these questions are usually backed by years of experience. AI initially has only the data it has been given. It can deliver excellent results – but only within the business context defined by the company.
In Food Wholesale, Experience Often Makes the Difference
Food wholesale in particular demonstrates why company knowledge matters so much. An ERP system such as Microsoft Dynamics 365 Business Central naturally contains extensive information: products, customers, suppliers, orders, inventory, batches, prices and much more. This data provides an excellent foundation. But it does not explain why certain decisions are made:
- A supplier may be preferred even though its price is not the lowest.
- A particular customer may order significantly larger quantities every summer.
- A product may deliberately be kept at higher inventory levels because replenishment at short notice is difficult.
There are good reasons behind all of these decisions. They come from experience.
The Most Valuable Knowledge Is Often Not in the ERP System
There is one thing I observe in almost every project: the better I get to know a company, the more often I discover that its most important information does not exist solely within the ERP system. It can be found in process descriptions, work instructions, Excel files, meeting notes or emails. And very often, it exists only in employees’ heads. There is nothing unusual about that. Quite the opposite. Many companies work so well precisely because their employees have years of experience and make decisions every day that cannot be fully documented. For AI, however, this knowledge remains invisible. Not because AI would be unable to understand it, but simply because it has no access to it. The goal is not to document every piece of experience down to the smallest detail. What matters is making visible the knowledge that regularly influences processes and decisions.
Why Many Companies Aren’t Ready for AI Yet – and Why That’s Actually Perfectly Fine
The good news is that companies are far from alone in this. Many projects reveal that data and knowledge are often treated as the same thing. In reality, most companies already possess a wealth of knowledge. It simply has not yet been structured in a way that AI can use reliably. Some of the challenges described here may have sounded familiar. If you have already implemented or modernised an ERP system such as Microsoft Dynamics 365 Business Central, you have successfully been through a similar process before. Back then, too, processes had to be understood, master data structured, terminology standardised and knowledge documented.
Those experiences now provide a valuable foundation for using AI. Being ready for AI therefore does not mean having perfect master data or documenting every single process in full. It means making visible the relationships, rules and experience that make your company what it is. In other words, making the knowledge that helps your business succeed every day accessible to AI. The foundation for this already exists in most companies. That is why it is perfectly fine not to be fully AI-ready today. AI readiness does not start with more data, but with a better understanding of your own business. And once you understand that, you may already be closer to your first successful AI project than you think.
