AI Product Management
Scoping around uncertainty, quality bars, offline-to-online metric ladders and shipping under model drift.
6concepts
65flashcards
42minutes of reading
- 01 Pricing and Packaging an AI Feature Why marginal cost changes the pricing question, the three models in use and what each fails at, and the guardrails a pricing decision needs when usage is heavy-tailed.
- 02 Prototyping AI Features Before the Model Is Ready How Wizard-of-Oz studies, model-in-the-loop prototypes and eval sets written as specifications let a team learn what users need and what the model must do before anyone commits to building it, and the specific ways each method lies.
- 03 Scoping Under Capability Uncertainty Why you cannot specify an AI feature the way you specify software, the cheap experiments that resolve the uncertainty, and the scoping decisions that determine whether a feature is buildable at all.