Add users, grow annual recurring revenue and expand multiples.
That model powered the rise of Salesforce, ServiceNow and hundreds of software-as-a-service (SaaS) companies whose valuations rested on predictable per-user subscriptions.
Artificial intelligence is beginning to fracture that logic.
As autonomous agents draft contracts, reconcile invoices, generate marketing copy and triage support tickets without tying activity to a named employee, the link between headcount and software revenue is weakening. Instead of charging per user, vendors are increasingly experimenting with pricing tied to tokens consumed, workflows executed, transactions processed, or measurable business outcomes delivered.
The shift represents a structural recalibration of SaaS economics rather than its demise.
Per-Seat Pricing Faces Structural Pressure
The per-seat model worked because it aligned incentives across vendors, customers and investors. Customers could forecast spend based on hiring plans. Vendors enjoyed recurring revenue visibility. Public markets rewarded expansion metrics and net retention rates built on seat growth.
But AI agents complicate that structure.
Software companies are reconsidering pricing as AI systems perform work that would previously have required multiple employees, the Financial Times reported Monday (Feb. 16). A customer support platform powered by AI, for example, may resolve a growing share of tickets autonomously. Charging per human support representative becomes less intuitive when much of the work is automated.
Per-seat pricing is not disappearing, but hybrid structures are gaining ground, Bain reported in October. Many vendors are layering AI surcharges on top of base subscriptions, charging additional fees for AI copilots or autonomous capabilities. Others are introducing usage tiers that scale with automation volume, such as the number of claims processed or documents analyzed.
The economic pressure is twofold. First, AI reduces marginal labor needs, weakening the historical correlation between customer hiring and software revenue. Second, AI inference carries real infrastructure costs, especially as models process large volumes of data. Vendors must design pricing that captures value without eroding margins.
From Licenses to Credits and Consumption
Bessemer Venture Partners’ AI pricing framework highlights a shift toward consumption-based models in AI-native firms, where revenue scales with API calls, tokens processed or compute cycles used.
Enterprise incumbents are adapting that logic. Credit-based systems are emerging, allowing customers to buy pools of AI capacity and allocate them across use cases. Transaction-based pricing is also gaining traction, with vendors charging per automated action, such as a generated marketing asset, reconciled payment or completed compliance review.
This structure more closely aligns price with output. If an AI system processes 100,000 invoices instead of 10,000, revenue scales accordingly. It also mirrors the cost base of AI infrastructure, where usage directly drives compute expense.
However, consumption pricing introduces volatility. Usage-based revenue can fluctuate with customer demand, potentially compressing valuation premiums that were built on predictable recurring revenue.
Outcome-Based Pricing and the Redefinition of Value
The most aggressive experiments move beyond usage and toward outcomes.
Some vendors are exploring contracts tied to performance metrics, such as faster loan approvals, higher eCommerce conversion rates or reduced fraud losses. Instead of charging for seats or API calls, they link fees to demonstrable business impact.
This model shifts risk. Vendors must prove that their AI delivers quantifiable gains. Contracts become more complex, and measurement frameworks grow critical. Yet in a budget environment where executives demand ROI clarity, outcome-linked pricing can resonate.
Per-seat subscriptions are unlikely to vanish overnight. Many enterprises still value the simplicity and predictability of user-based licensing. As AI agents scale and automate more complex tasks, however, the economic center of gravity is moving from access to output.
The SaaS era was built on selling licenses to people. The AI era is beginning to price around what the software actually does.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
