Test your app outside the golden path too

You have spent a few evenings building your app. Your agent says everything is ready and opens it in your browser. And then you click through it for a bit. The form works, the button works, the overview appears. It actually looks really good. You might recognise that moment, and also the feeling that you can publish now.

In the training on building with AI, I called that vibe checking. It is the sister of vibe coding. You are so in the flow that testing also happens a bit by feel. And then you very easily follow the ideal path, the golden path, as developers call it. What your app is supposed to do, it does. Only you do not yet know what it does when someone tries something else.

At the bottom of this article you will find two prompts to have your app tested thoroughly. Feel free to copy them.

Why your own test is too nice

I will give an example. Suppose you build an app that helps you to prospect. You have an ideal client in mind. And so you test with that same client every time, because that is the client you built the app for. That works every time. But what about a client without a website? Or with a name with an apostrophe? Or someone who half fills in the form and clicks back?

Those are the edge cases. They lie at the edge of what you want to do, and that is why you almost never encounter them yourself. You know how the app is intended. Your users do not know that. Ironically, you are therefore the worst placed tester of your own app, exactly because you know it so well.

Let your agent test itself, without repairing

The solution is actually very simple. You can test, but an AI can test too. You say to the agent you are building with: test it yourself now. Go through everything, click on everything, see what works and what does not.

Frequently asked questions

What is vibe checking?

Testing your app by feel, just as you built it by feel. You click through for a bit, everything seems to work, and without noticing it you only follow the ideal path. You do not test the edge cases then.

Why is the agent not allowed to solve the problems immediately?

An agent that repairs during testing jumps from one problem to the next. With a list you keep an overview and you tackle the issues one by one later.

Why let a second AI test?

Another model has different habits and did not help build. It therefore looks with different eyes. You give the found issues back to your original agent to solve them.

Where do I practise this with guidance?

In the Vibe Coding with AI training you build and test your own app, step by step. You can find the dates at Workshops & Events.

Keep reading for free

Register free with your e-mail address and read on. Your account opens the free articles in the learning space; the rest is for subscribers, from Learn Plus up.