Start small, then grow
Custom software does not have to start as a big project. I usually recommend building the smallest version that solves the most painful problem first, putting it in front of your team, and growing it from real feedback.
That keeps cost and risk under control, and it means you get something useful in weeks instead of months.