In the end it came down to either pressuring the client to embark on a complex cloud transition in time for a major sports event (that was a make or break event for the customer) or, as my friend recommended, start with a PoC but do the risky transition after the sporting event to minimize the potential for a catastrophic failure if things didn’t go to plan.
He didn’t get the job because even though the technical attributes were all good, he should have pressured the customer.
My takeaway was that interviews are a game with their own rules, and you have to play by those rules - demonstrating some sense of judgement outside the parameters of the game will not be tolerated.