Pricing the Machine: A Framework for Valuing AI Capabilities in B2B SaaS Business Models

Abstract

The integration of artificial intelligence into business-to-business software-as-a-service (B2B SaaS) is straining the pricing logic on which the industry was built. As value decouples from the number of people logging in, the per-seat subscription long the default no longer maps cleanly onto the value a product delivers, and vendors are migrating toward usage-, work-, and outcome-based structures. Yet the field lacks a rigorous, theoretically grounded method for answering the question these shifts raise: how much is an AI capability actually worth to a customer, and how should that worth translate into price? Drawing on two industry surveys of B2B software and AI companies (N = 240 in 2025 and N = 230 in 2026; Growth Unhinged/Tremont), this paper documents the empirical landscape hybrid pricing rising from 27% to 41% of firms within twelve months, seat-based pricing falling from 21% to 15%, AI gross margins clustering near 50% against SaaS norms of 70–80%, and a wave of AI-credit and outcome-based experimentation and identifies a gap between this fast-moving practice and the academic pricing literature. To bridge it, the paper develops the AI Economic Value Contribution (AI-EVC) framework, which extends classical economic-value-to-the-customer analysis with an explicit willingness-to-pay attenuation term, a value-capture parameter, a cost-and-margin floor, and a mapping from estimated value to hybrid pricing architecture. A worked example applied to an autonomous customer-support agent demonstrates the framework end to end and cross-checks its output against observed market prices. The paper closes with implications for pricing strategy, product development, and investor valuation, and with an agenda for empirical validation.