Before You Build That App

There is a moment happening right now in school networks across the country. Someone on the team discovers that AI can write software. Not technically write it, exactly, but close enough. A few prompts, a free trial, a weekend, and suddenly there is a working tool that does something your current systems never could. It feels like a breakthrough.

It might even be one. We are not here to tell you the moment is not real.

But we have seen enough of these stories play out to know that the beginning is almost never the problem.

Every school network has one. They do not have "technology" in their title. They might be a data coordinator, a program manager, or an operations lead. What they have is enthusiasm. They are the first person in the building to try a new tool, the one who figures out the workaround, the one who built the Airtable that everyone now depends on to track family outreach or staff onboarding or whatever problem needed solving at the time.

They built it because they cared. They maintain it because no one else knows how. And if they ever leave, the organization will spend months figuring out what they actually built and why it works the way it does.

This is not a criticism of that person. They are usually one of the best people in the building. It is a criticism of the setup.

Here is how internal tools tend to fail. Not all at once, and not right away.

First, no one plans for governance. The tool gets built to solve an immediate problem and it works, so it spreads. More people start using it. Someone adds a field. Someone changes a formula. No one is tracking any of it because that was never part of the plan. Eventually the tool is load-bearing infrastructure being maintained by one person in their spare time.

Second, that person becomes the single point of failure. When they are out, the tool is out. When they move on, institutional knowledge walks out with them. The organization is not left with a tool. It is left with a dependency it cannot explain.

Third, the tool hits a ceiling. The person who built it can only take it as far as their own skills allow. When the needs of the organization grow, the tool cannot grow with them. So another workaround gets built on top of the first one, and then another, until the whole thing is a structure that no one fully understands.

Now here is the honest part, because we think you deserve it.

The cost of building an internal tool has dropped significantly. AI-assisted development tools have made it possible for a motivated non-developer to build something functional in a fraction of the time it would have taken two years ago. Some of those tools are genuinely worth building. Simple, low-stakes, well-contained problems are reasonable candidates.

The cost of maintaining that tool has not changed at all.

Someone still has to own it. Someone still has to update it when the underlying platform changes. Someone still has to be available when it breaks on a Tuesday morning before a board meeting. And in most school networks, that someone is already doing another full-time job.

The question worth asking before you build is not "can we build this?" The answer to that question is almost always yes now.

The question is: "Who owns this in three years, and what does their week look like?"

If you cannot answer that clearly, you do not have a software plan. You have a prototype that is about to become someone's second job.

The good news is that there is a path that does not require you to choose between an expensive vendor contract and an internal tool your team cannot sustain. Lightweight, well-designed solutions built by an external partner, at a cost that works for social impact organizations, can give you what the internal build promised without the dependency it creates.

Sometimes the right answer is to build. Sometimes it is to design the data experience so that you do not need the custom tool at all. Knowing which is which is where the work actually starts.

If you are trying to figure out which side of that line your current situation falls on, we are happy to think through it with you. No pitch, just a conversation.