Almost every AI project starts with a finished solution in mind: an app, a chatbot, a dashboard. That is understandable - and often the reason why the project never becomes a good product.
In short
- The first idea usually describes the tool, not the problem.
- AI makes it cheaper than ever to build several variants - and to drop them again.
- Dropping early means investing later in what holds.
The first idea describes the tool
“We need an app” is a solution, not a question. Behind it is a task: finding something faster, backing up a decision, advising a customer better. Only once that task is clear can you judge which tool fits - and whether AI plays the main role at all.
That is why we do not start by asking “How do we build this?” but “How will your users notice that it got better?”
Dropping has become cheaper
Every variant used to be expensive: concept, design, delivery - trying three approaches meant three budgets. Today a clickable draft takes days instead of weeks with AI support. That changes the maths: it is cheaper to build three variants and drop two than to finish the wrong one.
Dropping is still work. It takes people who understand users, ask good questions and have the courage to let go of a nice idea. That is exactly where human creativity becomes more important, not less.
How we notice an idea does not hold
- Users do not need it - the prototype gets tried, but the task does not get easier.
- The data is not there - the answers would only be as good as sources that do not exist or nobody maintains.
- Operations break the business case - what works in the prototype would be too expensive, too slow or impossible to approve in everyday use.
We see the third point early because we come from delivery. We know what a prototype promises and what a product has to hold up to in operation.
An example: the request was an app
A manufacturer wanted a classic app for its sales team to look up equipment codes and prices - with a search field and filters. Instead we showed that a chat does the job more simply and built it as a working prototype: answers in seconds instead of minutes in spreadsheets. We did not build the app idea but dropped it - and created a better basis for the decision. To the case study
What this means for your project
Bring your idea - but do not hold on to it too tightly. In the reality-check sprint we test it together against variants, build the strongest one as a prototype and tell you honestly what is worth it.
