Time and materials, fixed price, retainer, outcome pricing: how each one moves risk, when it is the honest choice, and when it quietly becomes a trap.
A price is a sentence about risk. It says who pays when the work turns out to be harder than the room believed on the day of signature. On an AI project that sentence gets tested almost immediately, because a real part of the work is finding out whether the thing is possible at all.
Four models are actually used: time and materials, fixed price per deliverable, retainer, and outcome or value based pricing. The useful question is not which is cheapest, but which puts each risk on the side of the table that can see it, and what each rewards once the contract is signed.