Software projects result in digital products which solve a specific problem. These products will either be used by end users or other software systems.

Consider a scenario where a very complex AI model fuels a digital product which nobody needs. Evidently, it won't find traction among users.

Contrast this with a simple AI model powering a great product which is solving a real pain point. Exactly, users will love it.

Whenever I start a new project I thus try to get the bigger picture of why a certain problem needs to be solved and what the ideal product would look like.

I do this by talking to stakeholders and end users as well as by analyzing available data.

During this process it sometimes turns out that the problem can be solved without data & AI. Sometimes other problems emerge which are more important to solve.

What’s your 1st step when starting a new data & AI project?