Stefan Sarbu - 6 Apr 2026
How to build a business case for AI before writing a line of code
A pre-development business case for an AI initiative functions as a technical constraint document: the cost of what is displaced, the accuracy threshold required to sustain that displacement, and the maximum allowable cost per unit of output together determine architectural decisions before any model is evaluated. Without these constraints defined in advance, accuracy targets become intuitive, infrastructure choices become arbitrary, and the decision to move to production becomes a matter of engineering preference rather than measurable evidence. The sequencing discipline, business case before technical evaluation, is what makes success criteria testable and completion definable. Despite insane successes, we read about the vast majority of AI projects continue to fail. They do not because the technology stops working. They fail because nobody agreed, before a single line of code was written, on what "working" actually meant. This is a structural problem, not a technical one. Engineering teams begin evaluating models. Product teams scope features. Leadership approves a budget. And somewhere in that sequence, the foundational question gets deferred: what does this need to achieve, at what cost, to justify the investment?
