The question I hear most often is: how do I use AI in my business? Phrased that way, it almost always ends somewhere bad — because it starts from the tool instead of the problem.
The better question is: which task in my week takes the most time and requires the least judgement? Whatever that answer is, that is the right place to start.
Where it genuinely works
- Repetitive work with a clear input and output — summarising, sorting, pulling structured information out of text.
- The first draft of anything. Editing something mediocre is faster than starting from an empty page.
- Answering questions whose answers already exist somewhere and simply cannot be found.
Where it does not
Anywhere a mistake is expensive and hard to reverse. Anywhere the answer must be defensible rather than merely plausible. And anywhere you yourself do not know what a right answer looks like — because then you cannot tell a correct output from a well-dressed one.
A tool whose output you cannot judge is not a tool. It is a risk.
So good consulting is less about introducing tools than about drawing boundaries: these tasks yes, those no, and here is where a human still decides. Advice that hands you a list of tools has not given you time back — it has only made the job heavier.