Not every business problem needs custom software, and we'd rather tell a prospective client that a configured off-the-shelf tool solves their problem than sell a build they don't need.
Custom software tends to make sense when your process is genuinely non-standard — when the way you approve, route or report on work doesn't map cleanly onto what a generic tool assumes. If you're building workarounds around a SaaS tool's limitations, that's usually a signal.
It also makes sense when the software itself is part of your competitive advantage — when how you serve customers or run operations is meaningfully different from your competitors, and a generic tool would flatten that difference.
It tends not to make sense when the problem is well-solved by existing tools and your workflow isn't unusual enough to justify the build and maintenance cost. In those cases, the honest recommendation is to configure what already exists.
A useful test: if you can describe your process in one sentence and it sounds like every other business's process, buy. If describing it takes a paragraph of exceptions and special cases, that's usually where custom software starts to pay for itself.