Fifteen years ago my staff stood at my desk. The servers were moving to the cloud. Their question: “Tim, will my job still exist in five years?”
I wrote a book about it back then: Cloudsource je ICT!.
Now the same question is back on the table. Only this time it is about AI. And about a new promise: vibe coding. You tell the AI what you want and the AI writes the software. So why would you still buy standard software?
I get the question. We use AI heavily ourselves. Still, I do not believe for a second that standard software is going away. My answer fits in two sentences:
Buy what everyone has. Build what nobody has.
And let AI help you with both.
Guess who is still here in five years
In my company I now see exactly what I saw fifteen years ago. Especially among developers. One group cries bloody murder. The other dives in and tries everything.
Guess which of the two will still be here in a few years.
That is how it went with the cloud. The system administrators who moved along became cloud engineers. Those who clung to their server rack fell behind. The work stayed. It just took a different shape.
AI will go the same way. That is why I look at vibe coding with peace of mind. And with a critical eye at everyone shouting that standard software is dead.
Three questions before you build it yourself
Thinking about building your business software yourself with AI? Ask yourself three questions first. Can you do it? Will it keep working? And is it smart? Let me walk you through them one by one.
Question 1: Can you do it?
Every revolution brings new complexity. A big technological leap makes things easier at first. Then the complexity arrives. And that calls for its own specialists. Look at your network. As a business owner you could perfectly well maintain it yourself. Yet you outsource it. You want to spend your time on your business. Vibe coding works the same way. You became an entrepreneur to run a business. Do you really want to learn a whole new trade on the side?
A proof of concept is not business software. You build a working proof of concept in a weekend. Software your business runs on takes months. Handling exceptions, user rights, speed, backups and an audit trail for your accountant. That is the last 20%. And it takes 80% of the time. You never see that part in the videos on LinkedIn.
Building it yourself means testing it yourself. After building comes testing. Catching errors. Working out why that invoice went out twice. And there is always that one colleague who understands how it works. If he gets sick or leaves, you are stuck. That exact dependency is why so many companies once moved to standard software. Vibe coding brings that risk straight back in. Only now it is called innovation.
Question 2: Will it keep working?
Say your app is live. It works. Great. That is when it really starts.
Code ages. Code nobody truly understands turns into technical debt. Components go out of date. Security holes are discovered. Who patches your home-built app three years from now? With standard software the vendor does that. For all customers at once.
Security is a profession of its own. And it is growing faster than almost any other. Application hardening, access management, encryption, penetration testing, monitoring. More gets added every year. And AI makes hackers faster too. A serious software vendor has a dedicated team that does nothing else. How many SMEs can say the same? If you build it yourself, you are responsible for its security yourself. Including all the customer data and business information that comes with it. As a business owner, that is not something I want to carry.
The law changes, and your software has to follow. VAT rules change. E-invoicing becomes mandatory. GDPR, NIS2, Peppol. Standard software handles it with an update. Your own app does not do that by itself. I do not know a single business owner who wants to lose sleep over that.
Question 3: Is it smart?
Say you answer yes to the first two questions. Then the most important one remains. Is it wise?
You get the experience of thousands of companies. Many SMEs actually want guidance. They want to work with processes that have already proven themselves elsewhere. Reinventing the wheel? They do not have the time. They would rather spend it on their core business. I could not agree more. Standard software delivers that knowledge out of the box.
Sharing is cheaper than paying alone. Say developing your own app costs less than 1% of a standard package. You are still paying too much. The development costs of standard software are shared by hundreds of thousands of customers. You pay the bill on your own. Your competitor splits it with a hundred thousand others. And I have not even mentioned hosting, security, your own hours and the hours of that one colleague. Work out the total cost and the gap only grows.
The ecosystem sets you free. Standard software makes you less dependent. Including on your implementation partner. Yes, I am saying that as an implementation partner. There are standard integrations, documentation, support, a community and other partners. Not happy with us? Then you switch. That is how it should be. Try that with custom code only your former colleague understands.
Standard software is getting smarter too. The vendors are not standing still. Odoo and other packages are building AI into their products. That only widens the lead of standard software. Vibe coding will mostly happen on top of a platform. More on that in a moment.
Buy what everyone has, build what nobody has
So is building it yourself never smart? It is. There is one clear exception.
Do you have a process customers choose you for? Something your competitor cannot do? Then build that yourself. With AI, that is now affordable too. That is where your edge lives, and you want to keep it in your own hands.
For everything else: buy it. Your accounting, your inventory, your purchasing. Nobody becomes your customer because of those. Those processes just need to run well. Every day.
What does change
So standard software stays. Does that mean nothing changes? Quite the opposite. A lot is changing. And honestly, that gets me excited.
Implementation gets faster. A solid inventory of what you really need is still the foundation. Without it you build on sand. But AI makes that discovery much faster. Collecting and interpreting the questionnaires and turning them into requirements used to take us 8 hours. Now it takes an hour at most. We put those 7 hours into what really counts: sitting down with the customer, digging deeper, or preparing a proof of concept.
Learning the system gets easier. People still need to learn new software. But well-documented standard software can now simply be asked questions through AI. How do I create a credit note? Where are my open orders? You get the answer right away.
Small extensions cost next to nothing. This is where vibe coding really shines. An automation, a dashboard, a handy little tool on top of your standard package. You now build it in a fraction of the time. That slashes development costs. A proven foundation with smart custom work on top. To me, that is the ideal combination.
Getting people on board is still human work. AI speeds up a lot. But no tool gets people to work differently for you. That is why the role of the implementation partner is shifting. Less configuring. More advising, guiding and bringing people along.
Who should be afraid?
So AI is mostly good news. Still, there are four groups I would warn.
1. The business owner who stays put. He clings to his old system. It works, right? Meanwhile his competitor works faster, smarter and cheaper. Every year he falls a little further behind. I find that a waste. Because it is so unnecessary.
2. The organisation that bets everything on AI blindly. The other extreme. They chase every self-proclaimed AI specialist. They pour time and money into experiments that never go live. I recognise some of those specialists, by the way. A few years ago they were selling courses on getting rich by sourcing from China. Then it was crypto and NFTs. Now they sell AI.
3. The vendor who keeps the gains. AI makes software and implementations cheaper. I believe those gains belong with the customer. In lower prices, fewer implementation hours or more value. That is why we now use AI in our discovery work. Vendors who keep charging their old rates and hours will lose customers to those who do pass it on.
4. The partner who sells hours. Does your business model depend on selling as many hours as possible? Then you have a problem. Customers increasingly pay for results. Rightly so, if you ask me.
My advice
What should you as a business owner do with this? Three things.
- Choose a standard package as your foundation. For everything every business needs.
- Only build what sets you apart. AI can help you plenty there.
- Ask your vendor what AI gets you. Lower costs, fewer hours or more value. No clear answer? Then you know enough.
And now the good news
Because it is not all doom and gloom. In my own company, AI has made things possible that were out of reach for a team of fourteen.
Take our CPQ Builder. For years, customers asked for a product configurator that builds quotes and orders automatically. It existed, but only for large companies with large budgets. There was no affordable alternative for SMEs. With AI we built one ourselves. Exactly the kind of custom work I described above: something nobody had, on top of a standard package.
I wish that on every entrepreneur. So yes, standard software stays. And AI makes it better, faster and cheaper. Both are true.
Fifteen years ago my staff asked whether their jobs would survive. The work is still here. It just looks different. AI will go exactly the same way. The software stays. The way we build it, sell it and use it is changing.
So buy what everyone has. Build what nobody has. And start today.
Read more: Custom software just got affordable for mid-sized companies · In three months, AI changed our company · How a fit-gap analysis saves your project