Findings

The AI software market is shifting away from seat-based subscriptions toward consumption-based pricing models - credits, tokens, usage tiers, and outcome-based fees - under the rationale that AI's variable cost structure requires a different commercial approach than traditional SaaS. This study examines how that shift is being received by the buy side.

Benchmarkit, in partnership with Pricing I/O, surveyed 296 software buyers between February and April 2026. 

Respondents ranged from companies under $20M in revenue to enterprises above $10B, and 90% held Director-level titles or above. The survey examined four areas: how buyers evaluate AI pricing pre-contract, how AI spend has performed against budget, which pricing models buyers

prefer, and how AI pricing is perceived relative to traditional SaaS. The findings indicate that buyer sentiment is not aligned with the direction the market is moving.