Post-mortems love to blame technology. But rewind any failed platform project and the fatal decisions cluster in the same fortnight — the one where everyone was eager to 'get started' and nobody wanted to slow down for questions.
The questions that don't get asked
Who actually performs this process today, and what do they do when the rules don't fit? Which of these requirements is a legal constraint versus a habit? What happens in March, at year-end, during the audit? The exceptions are the specification; the happy path is just its cover page.
Requirements as testable statements
'The system should handle returns efficiently' is a wish. 'A return against a partially-paid invoice creates a credit note and adjusts the ledger within the same day's books' is a requirement — you can build to it, test against it, and argue about it before it costs anything.
Spike the risky assumptions
Every blueprint has two or three assumptions carrying disproportionate risk: the legacy integration, the data quality, the AI accuracy. Prototype exactly those in week two, not month eight. A ₹2 lakh spike that kills a ₹2 crore mistake is the best trade in software.
Discovery isn't a delay before the real work. It is the real work, done while changing course still costs a meeting.