Does the citizen developer exist? Revisited.

Back in 2020 I asked a simple question: does the citizen developer exist? My answer was a cautious yes. Six years, one AI wave and quite a few projects later, it’s time to check my own homework. Spoiler: I got the destination roughly right, and the route mostly wrong.

The question originally popped up when I was helping a company build some simple internal process flows. They wanted to build everything themselves, but needed some help to get started. From there on, they went on to digitize some other simple processes, all on their own. The fact that they couldn’t start on their own raised the question: does the citizen developer exist?

What I said in 2020

In the original piece, I argued that low-code and no-code platforms would let business users build a real share of their own software. The people who use an application every day would increasingly be the ones who build it.

I also warned that it wouldn’t be easy. Simple apps, sure. But as soon as you needed to connect to Active Directory, talk to another system over a REST API, or design a data model with a few dozen tables, you were no longer in drag-and-drop territory. Those technical hurdles, I thought, would decide how far citizen developers could go, and those hurdles appear as soon as a larger part of the organization gets involved.

Sounds reasonable, right? Well… partly.

What actually happened

The number of citizen developers grew. That part I got right. But ask how many there are, and things get vague quickly.

Reports talk about tens of millions of citizen developers. Look closer and those numbers are mostly forecasts. Gartner predicted (opens in new window) that citizen developers would outnumber professional developers four to one by 2023. A bold prediction, but I haven’t seen anyone actually count them since. What is measured are developers in general, and even there the estimates range from 27 million (opens in new window) to 47 million (opens in new window) worldwide, depending on who counts and how.

And the number of citizen developers isn’t even the interesting question. The interesting question is: what did they actually build? How many of those applications went into production, without the help of professional developers or IT, and are still being used and maintained today? As far as I can tell, nobody counts that.

What I have seen is a shift in tools. Business users who would have reached for a no-code or low-code platform a few years ago now increasingly use AI-assisted development. And what they build is mostly small: automated workflows, handy additions to Excel and PowerPoint, tools that make one person or one team faster. Big systems built by citizen developers? I’ve seen almost none.

Where they really got stuck

Here’s where I was wrong. Yes, integrations and data models tripped up some citizen developers. But that wasn’t the real wall. The real wall was governance. Or rather: the lack of it.

In practice, that looked like this:

• Security was rarely considered, let alone tested. Which makes an app very hard to roll out across an organization.
• Deployment beyond the maker’s own desk. Building an app is one thing. Getting it to the rest of the company without the right governance in place is something else entirely.
• Access control. Most companies have clear rules about who may see what. The citizen developer usually doesn’t know them, or doesn’t know how to comply without serious help from IT.
• Data storage. The same story: companies have rules about where data may live, and citizen developers often aren’t aware of them or lack the knowledge and support to follow them.
  And I’m not even talking about backups…

Two examples I’ve seen up close. At one company, citizen developers built applications that genuinely boosted their productivity. Then they had to stop using them, because the apps didn’t meet the organization’s security standards and access controls.

At another company, a whole collection of Excel sheets was replaced by citizen-built apps. Great result, until IT found out. The data was now stored outside the organization, which was very much against the data governance rules. The apps were taken down, and the team went back to their old Excel sheets. Productivity gains: gone.

And even when security, access and storage were in order, there was the question nobody wanted to answer: who supports this? IT, who knows nothing about these apps? Or the business, who has neither the time nor the skills?

So the pattern I see is this: lots of small apps. They help one employee, sometimes a whole team. And that’s where it stops. Sometimes because of poor testing or quality, but even when all of that is fine, governance makes it really hard to roll citizen-built apps out across a company.
An entire new level of shadow IT was created.

Then AI arrived

AI-assisted development removes even more of the technical barriers. You describe what you want, and you get something that works. Or at least something that looks like it works.

That sounds like the citizen developer’s dream. But it doesn’t solve the governance problem. It adds new questions to it:

• Is the generated logic correct? If nobody checks what the AI produced, nobody knows.
• Is the AI using an approved model, and is data stored according to the company’s rules?
• Can the output contain confidential information? Data and privacy don’t stop being an issue because a model wrote the code.
• Are users relying on an AI answer without verifying it? This question comes up every single time.
• Are decisions auditable and explainable? Citizen developers are often unaware this is even a requirement, let alone how to meet it.
• Who is accountable when an AI-generated workflow causes an error? IT? The business? Because there’s often no check on what the AI generated, the answer in practice is: finger-pointing.

So the wall didn’t disappear. It moved. And with AI making it so easy to build, more people reach that wall, faster, while the other governance issues still remain.

So, does the citizen developer exist?

Yes. More of them than ever, and AI is only accelerating that. But the citizen developer I imagined in 2020, building business applications that spread across the organization, mostly didn’t happen. Not because the technology fell short, but because the organization around it wasn’t ready.

Where bigger adoption does happen, it’s because governance is in place and citizen developers work strictly within it.

That also shows in the language. “Citizen development” is still widely used and understood, especially by analysts, vendors and governance teams. But in their own operating models, many organizations now prefer terms like business technologist, maker or fusion team. They sound more professional, and they better reflect what actually works: business and IT building together, within agreed rules.

So a 2026 version of my 2020 answer might read: a company runs a governed citizen development program, with business technologists or makers working in fusion teams, on approved low-code platforms or AI tools.

What to do now

If you’re a CIO, a development manager or a business owner wondering what to do with all this, here’s where I’d start:

1. Just get started. Don’t wait for the perfect framework. Try citizen development on a small scale and see if it works for you. Find a sponsor inside your organization, and find out whether the organization is actually willing to adopt it.
2. If it is, set up governance before you scale. Agree where data may live, which access rules apply, which platforms and AI models are approved, and how AI-generated logic gets checked. Make it easy for citizen developers to comply, not just possible.
3. Decide on support and accountability up front. Who supports a citizen-built app once more people use it? Who is accountable when something goes wrong? Answer that before the rollout, not after the first incident.

Do that, and your citizen developers, business technologists, makers, or whatever you call them, get the chance to build something that lasts beyond their own desk.