Your business is not a guinea pig. Stop treating it like one!
AI is a great tool. We use it. We recommend it. It saves time, cuts errors, and handles tasks that would have taken hours just a few years ago.
But there is a version of AI use that is quietly causing real damage to SMEs running Microsoft Business Central. Not through dramatic failure. Through small, confident mistakes that go unnoticed until they are expensive to fix.
This is not an argument against AI. It is an argument for knowing where it stops being useful.
When Fast Becomes Costly
SMEs running Business Central are lean on IT resources and short on time. When something goes wrong, the instinct is to fix it fast. Nobody wants to log a support call. Nobody wants to admit they are not sure what to do. So they type the problem into an AI tool and follow the steps it gives them.
Or they ask AI to write code. Build a customisation. Automate a process. The AI generates something that looks professional, finished, and correct. And the logic makes sense: if AI can write the code, you save the cost of a developer.
That saving is real. But it is also temporary.
Writing code is only part of the job. You also need to understand what that code does, how it fits into your system, and what to do when something goes wrong. That last part is where the cost starts to climb.
Businesses that go down this route tend to encounter the same pattern. An issue appears in Business Central. Rather than calling their partner, they ask an AI tool for the answer. They follow the instructions. The issue appears to be resolved.
It is not resolved. It has moved. By the time they realise, the knock-on effects have spread through several months of data. What would have been a straightforward fix becomes a much larger undertaking. More time. More cost. More stress for the team.
The same pattern appears with code. A business asks an AI tool to generate code for a custom report. It looks good. It runs. But the logic is subtly wrong. For months, the management team has made decisions based on slightly incorrect figures. Nobody knows until someone starts digging deeper. The cost of fixing it, in terms of both labour and decisions made based on bad data, is high. Far more than the developer cost they avoided in the first place.
Why AI Gets It Wrong in Business Central
AI tools are trained on general information. They can explain how Business Central works at a conceptual level. They can point you toward menus and help you understand terminology. What they cannot do is see your system.
They do not know how your chart of accounts is structured. They do not know which customisations your partner built. They may not know that your stock module is linked to your sales platform. They may not know your VAT setup was changed for a specific reason two years ago.
When AI tells you to change a posting group or adjust a dimension setup, it is giving you a generic answer. That answer might work on a default Business Central installation. It might break yours.
There is also a confidence problem. AI gives answers in a clear, authoritative tone. No warning label. No caveat saying “this advice is based on a standard configuration and may not apply to your environment.” Just clear steps that feel correct. If you are not a Business Central expert, you have no way of knowing whether that advice fits your situation or not.
The same applies to generated code. AI can write code that looks professional and runs without errors. That does not mean it is correct. It does not mean it handles edge cases. It does not mean it integrates properly with your existing data structures or will hold up as your data volumes grow.
Understanding what code does, why it does it, and how to fix it when it breaks. That is where the real cost lives. That is expertise you cannot replace with the tool that wrote the code in the first place.
The Specific Risks for Growing Businesses
SMEs have layers of complexity that make this riskier than it might appear.
Many handle multiple revenue streams. Wholesale, retail, services, recurring contracts. Business Central needs precise configuration to track these separately and report on them accurately. A wrong adjustment to a posting group can have consequences across your whole ledger. Code that does not account for these distinctions can cause months of reporting problems.
Integrations are often critical. Your Business Central might be connected to your e-commerce platform, your CRM, your payment system, your warehouse management software. If the posting logic gets disturbed, data flows break. If custom code does not handle the data formats it receives, records fail to sync. Your teams start working with stale information. Your processes break.
Stock management, multi-currency trading, and complex project costing add further layers. These are not edge cases. They are day-to-day realities for growing businesses.
You can reduce the upfront cost of development by using AI to write code. But the cost of understanding that code, checking if it is correct, and fixing it when something goes wrong remains. It just moves further down the line, and it tends to arrive at the worst possible moment.
What to Do Instead…
Use AI. It is genuinely useful for loads of things.
Use it to draft emails, summarise documents, look up what a Business Central feature does at a conceptual level, or plan a project.
Do not use it to diagnose and fix live system issues. Do not use it to generate code for your Business Central customisations. Have that code reviewed by someone who understands your system. Make sure that a person can be held accountable for it.
If something is not working in Business Central, call your partner. If you need custom code, talk to your partner. A good Business Central partner knows your setup. They have your configuration documented. They can see what others cannot. They understand the consequences of changes. And they can be held accountable if something goes wrong.
A support call or development conversation might feel like an expense. It is almost always cheaper than trying to solve it yourself with AI. It also avoids the risk of finding out six months later that your data is wrong.
If support costs are a concern, ask about a support plan. A set number of hours each month means you are not weighing up whether to call for help against the budget. You just call.
AI Is a Tool. Use It Like One.
Every tool has a job it is suited for and jobs it is not. A good tradesperson knows the difference.
For system issues and custom code in a complex ERP like Business Central, the right resource is a qualified person who knows your system. Not a language model working from general training data.
Your business is not a test environment. The data in your system is real. The decisions your team makes from that data are real. The cost of getting it wrong is real.
Get in touch with Tecvia before a small problem becomes a large one.
FAQs
For anything not covered here, get in touch directly. We’re happy to answer questions specific to your business and your ERP requirements.
AI can help you understand what a feature does or where to find a setting. It cannot see your system, your configuration, or your data history. If something is wrong in your Business Central, contact your partner. A generic AI answer applied to a specific setup can make the problem worse.
AI can generate code that looks correct and runs without errors. That does not mean it is right for your system. Business Central customisations need to account for your existing data structures, integrations, and business rules. Code that ignores those factors can cause reporting errors, data issues, and process failures that are expensive to unpick.
The upfront saving is real, but it is often temporary. Writing the code is only part of the cost. Understanding what it does, maintaining it, and fixing it when something goes wrong requires expertise that AI cannot provide. In many cases, businesses spend more resolving AI-generated code issues than they would have spent on a qualified developer from the start.
Call your partner. The sooner you do, the smaller the problem tends to stay. Self-solving with AI can turn a quick fix into a weeks-long recovery job. Your partner knows your setup and can diagnose the issue properly.
Yes. A Tecvia support plan gives you a set number of hours each month, so you are not weighing up the cost of a support call every time something comes up. Visit tecvia.co.uk to find out more or book a free consultation.
Business Central is a complex, interconnected ERP system. A change in one area can affect postings, reporting, and data flows across the whole system. AI tools are trained on general information and have no visibility of your specific configuration. That makes generic AI advice particularly risky in a Business Central environment compared to simpler, more standardised software.
