The question everyone avoids
If AI writes the code, infrastructure is a platform, and maintenance becomes a prompt: who is actually in the driver's seat? The first article in this series argued that large IT departments have had their best days. But that raises a logical follow-up question: should business then start building its own apps?
The answer is nuanced. And it begins with a story.
A project under maximum pressure
Years ago, I worked on a project for Carrefour. The time pressure was extreme. The kind of pressure where there's no room for slow communication, lengthy meeting cycles, or weeks of waiting for feedback. We had to do things differently.
What we organised was, in essence, very simple.
Every Monday before 9 AM, a new build was ready on the test environment. Carrefour's business owner spent the entire day testing what had been delivered the previous week, together with the development team lead. Bugs were dispatched immediately. The team started fixing them right away.
Monday afternoon or Tuesday morning, there was a discussion about what needed to be delivered that week, alongside the bug fixes. That was passed on directly to the team, work was distributed, and everyone got to it.
Wednesday, the people from Carrefour came on-site. They reviewed what was already built together with the developers. Direct alignment, no intermediary layers.
Thursday, the team continued working.
Friday, the business owner brought in a full testing team. Everything delivered that week was thoroughly reviewed, and developers were corrected on the spot where needed.
Monday, the cycle started again.
It was a lot of pressure. Honestly: too much pressure to sustain. But in my entire career, I have never worked on a project that performed better than this one. Never. The reason was simple: business and development were so close together that there was barely any room for misunderstandings, wrong priorities, or lost context. Feedback was measured in hours, not weeks.
That is exactly where software development needs to go
What that Carrefour project put into practice was something agile had long promised but rarely delivered: genuine collaboration between the people who understand the problem and the people who build the solution.
The problem with the classic approach is not that developers are bad. It's that the distance is too great. Business writes a requirements document. That goes to IT. IT interprets, builds, delivers. Business sees the result weeks later and only then discovers that half of it is wrong. The iteration cycle is too long to course-correct quickly.
AI changes that dynamic fundamentally, but not in the way most people expect.
It's not about business sitting alone behind a screen typing prompts until an app comes out. It's about a new form of collaboration: a more technical person as an assistant to a business analyst, a marketer, a UI expert, a support employee — people who know the problem from the inside. Together, they start building with AI. In a day's time, they put an application together. They test together. They adjust. And a platform handles the deployment immediately.
What used to take weeks now takes a day. What used to be a months-long cycle becomes a week.
The Carrefour approach was, at the time, the maximum possible with the tools we had. Today, that maximum has moved considerably further.
For smaller apps: business works alone
Not every app even requires that tandem. A marketer who wants a tool to summarise campaign results. A support employee who wants to build an internal FAQ system. An account manager who wants to automate quotes.
For those use cases, business can already largely work alone today. AI guides the process, the platform handles the rest, and the result is live without a single developer ever getting involved.
That sounds radical. But it's not so different from what happened when the smartphone arrived.
We've been here before
Remember the fear that gripped IT organisations when smartphones and tablets started appearing in the workplace? Bring Your Own Device (BYOD) was the future and simultaneously the nightmare of every IT department.
Employees wanted to use their own devices. IT saw security risks, management complexity, an uncontrollable proliferation of devices on the network. Endless policies were written. Some organisations banned BYOD for years.
Today, it's completely normal. Nobody gives it a second thought. The tools have matured, security practices have kept pace, and the added value — people working on the devices they know and find comfortable — has completely overshadowed the initial fear.
With AI coding and business building on its own, we are at exactly the same point. The fear is understandable. The objections are partly valid. But the direction is irreversible. In five years, nobody will be surprised that a marketer built her own tool.
What this requires of organisations
The shift is not purely technical. It also demands something from how organisations think about roles, responsibility, and collaboration.
A business analyst who builds together with AI has not become a developer. She has remained a business analyst who now has a more powerful tool. But she does need someone to help her over the hurdles: someone who understands the technical context, knows when something is more complex than it appears, and ensures the whole thing stays solid.
That is the new role of the technical expert: not the builder, but the guide. Not the department that controls everything, but the person who sits at the table at the right moment.
Just as the Carrefour business owner was present on Wednesdays and Fridays — not to control, but to look together, course-correct, and confirm that things were on track.
The circle is complete
Linking this back to the first article: you bring in the expert when you need them. For the smaller app, that may be never. For the more complex application, that might be one day per sprint, or a code review at the end. For the app with sensitive data, that's a targeted security audit.
A platform that facilitates all of this — building, deploying, and finding the right expert when needed — is not the future. It is the logical conclusion of a shift already underway. DotBots Boutique is one of the platforms heading in that direction: infrastructure and security built in, peripheral processes supported, and matching with the right expert when you need one.
Business builds. Experts guide. Platforms handle the rest.
That is the new model. And just like BYOD, in a few years it will simply be normal.
Last updated: 2026-06-04 11:35