I want to say that you don't "ensure a business case truly reflects the expected benefits before starting a project" Technically, you could write it to say whatever you want, but it's best if you ensure it is credible, transparent, decision-ready, and defined well-enough to track over time. You do this by:
- Anchoring benefits in reality and validate them throughout the project
- Making assumptions explicit - Defining measurable outcomes
- Pressure-testing the assumptions - challenge them to see if they hold up and identify which would break the business case if they don't hold up
- Creating alignment on the meaning of success