"Let's do it with AI" has become a phrase said far too often, but not every problem needs a custom technical solution, and not every custom solution needs AI in the first place. Before investing in a custom solution, here are questions worth asking yourself first.
Is the problem actually clear? "I want something to make my work easier" isn't a specific problem. "I want to automate manual order logging that takes an hour daily" is a clear problem that can actually be solved. The more precise the problem, the clearer the solution.
Does the benefit justify the cost? A custom solution takes more time and cost than an off-the-shelf tool. If the problem happens once every couple of months and takes ten minutes, it probably doesn't justify a custom solution. If it happens daily and takes hours, the math changes.
Is the necessary data actually available? Any data-driven solution (especially one involving AI) needs sufficient, reliable data first. Without enough data, the "solution" becomes a guess wrapped in a modern interface.
Have you tried ready-made tools first? Many "problems" are already solved by existing off-the-shelf tools at a much lower cost than building from scratch. A custom solution is only worth it when you genuinely need something no ready-made tool provides.
If after these questions the problem still feels clear and the benefit real, that's the right time to discuss it together. We honestly study the problem's clarity and feasibility before any commitment — including an honest assessment of whether AI actually serves your solution or is just an unnecessary addition.