Engineering pricing, packaging, and unit economics for AI-native products where token cost, hallucination risk, and switching cost shape margin in non-obvious w
AI business model design is monetization craft for the inference era: pricing against value when marginal cost is tokens, choosing seat vs. usage vs. outcome models, defending margins as model costs and capabilities shift, and designing data flywheels that compound defensibility.
Founder/CPO-critical: investors probe AI unit economics hard in 2026, and operators who design margin-resilient models raise easier and survive pricing shocks competitors don't.
Increasingly hybrid: predictable base (seats/platform) plus usage or outcome components aligned to value and inference cost. Pure seats leak margin under heavy use; pure usage scares buyers: design for both truths.
Engineer the COGS (routing, caching, small models), price above cost-to-serve with headroom, and build value layers (workflow, data) that justify price when raw capability commoditizes.