Start with the constraint, not the technology
Custom software is worth considering when an important workflow is repeatedly forced around the limits of generic tools. The strongest cases usually involve connected processes, specialised knowledge, or a customer experience that is central to the business.
It is not automatically the right answer. If a proven product handles the need well, configuring it may be the simpler choice. A useful discovery process should make that visible early.
Look for durable value
The question is whether a tailored product can remove a meaningful operational constraint or enable something the business cannot do well today. That value should be understandable without relying on ambitious forecasts.
Begin with the essential workflow, define what success means in practical terms, and build enough to learn before expanding the scope.