The questions a product leaves unanswered

Products have a way of growing around recurring problems. We add an email to explain a confusing step, a check to catch a common mistake, a bot to answer a question support sees every day. Each addition can help. It can also make the underlying problem easier to live with, and therefore easier to leave in place.

A repeated support question is worth looking at before automating the answer. Is the information missing, buried, or arriving too late? Sometimes the most useful change is to make a status clearer, put an answer where it is needed, or remove the step that prompts the question in the first place.

The difficult call is knowing when to stop patching one piece at a time. Rethinking a flow costs time now; keeping exceptions, messages, and workarounds costs time every day. If the fixes keep piling up around the same problem, when does it make more sense to rethink the flow than to add one more?