How do I AI? Part one: understand what you're working with
Part of a short series on approaching AI in a business. The previous piece covered getting the systems underneath into shape. This one is the groundwork before you build anything: models, cost, privacy and where your data goes.
"We want to use AI" is something we hear quite a lot.
The problem is that it's not specific enough. It's a bit like saying "we want to use software". The technology might be useful, but until you understand what you're trying to achieve, it's difficult to know what the right implementation looks like.
For some businesses the first useful step might be using ChatGPT or Claude more effectively. For others it could be automating part of an existing process, connecting AI into business systems, processing large amounts of information, or building something entirely new. Sometimes, after looking at the problem thoroughly, AI won't be the answer at all.
So before you start building anything, there are a few things worth understanding.