Bespoke software is justified when an important requirement remains unmet after the sensible existing routes have been examined. The usual order is: use an existing product, configure it, integrate it, automate around it, extend it, and only then consider a fully bespoke system.
The decision should account for more than build cost. Custom software creates an ongoing responsibility for security, hosting, support, change, documentation and eventual replacement. A product subscription can be the better engineering choice even when building looks possible.
Discovery should produce a written comparison: the process and users, required outcomes, data and integrations, available products, important gaps, risks and the cost of operating each option. If a build is warranted, those findings become the basis for scope and acceptance—not a promise made from a short feature list.