From Generative AI Capability to Business Value: A Critical Review of Verification Complements and Value Capture

Generative artificial intelligence can reduce the cost of producing text, software, and other digital outputs, yet improved task performance does not necessarily produce a viable business model. This review examines the mechanisms connecting generative capabilities to value creation, delivery, and capture. A purposive synthesis brings business-model and digital-transformation theory into dialogue with experimental productivity and creativity studies and recent business-use-case research. The paper distinguishes operating improvements within an existing model from changes to the customer promise, activity system, or revenue logic. It develops a conceptual framework in which useful generation creates potential value, complementary assets enable delivery, verification determines acceptable quality, and bargaining conditions influence who captures the resulting surplus. Four business-model archetypes are examined through their revenue mechanisms, necessary complements, and failure conditions. An explicit cost-per-accepted-outcome model and a hypothetical service example show how inference savings can be offset by review, rework, and provider dependence. Four testable propositions and explicit cost thresholds connect task verifiability, complementary assets, differentiation, and supplier portability to business outcomes. No new field data, causal estimates, or commercial forecasts are reported. The review concludes that research should measure accepted outcomes and distribution of value across stakeholders rather than infer business-model innovation from tool adoption, usage growth, or faster content production alone.