KeyBanc Warns: AI Disruption Looms for Software Revenue Models, 'Activity-Based' Models Most Resilient

A new analysis from investment bank KeyBanc Capital Markets delivers a stark warning to the software industry: not all revenue models are created equal in the face of artificial intelligence. The firm argues that the traditional, widespread practice of charging customers based on the number of user 'seats' or licenses is now highly vulnerable to AI-driven disruption. This model, a cornerstone of enterprise software for decades, faces direct pressure as AI tools automate tasks and potentially reduce the number of human users required to perform the same volume of work.

KeyBanc's research suggests that software companies whose revenue is directly tied to measurable customer activity or usage—such as transaction volume, data processed, or compute cycles consumed—are in a far stronger defensive position. This 'activity-based' pricing aligns the vendor's success with the customer's operational throughput, which AI is more likely to augment than diminish. In essence, as AI makes a customer's business more efficient and productive, the volume of underlying activity (and thus software revenue) could grow, creating a natural hedge against displacement.

The implications are significant for investors and software executives evaluating long-term strategy. The warning signals a coming period of scrutiny and potential repricing for companies reliant on per-seat licensing, especially in productivity, CRM, and collaboration sectors where AI automation is rapidly advancing. It pressures management teams to justify their pricing architecture's durability and may accelerate a strategic pivot towards usage-based and outcome-based commercial models to future-proof their businesses.