Software vendors are quietly killing the per-seat subscription model. In its place, they are rolling out granular, usage-based billing that treats AI agents like utility workers. HubSpot now charges $0.50 per resolved conversation, down from $1.00. Salesforce has introduced Agentforce Flex Credits, billing $0.10 per standard action and $0.15 per voice action. Zendesk is charging $1.50 to $2.00 per automated resolution, while Sierra AI commands a premium, with third-party estimates placing their fees between $1 and $2.50 per resolved outcome. ServiceNow has also moved to a consumption-based model that bills for each AI skill execution, or ‘assist.’
The per-seat SaaS model worked for decades because it offered predictable revenue for vendors and predictable costs for buyers. That logic falls apart when software stops being a tool for a human and starts acting as the worker itself. If an AI agent handles a customer service ticket, there is no longer a human ‘seat’ to license. Vendors are scrambling to capture the value of that automation, leading to a messy, experimental phase where outcome-based, consumption-based, and hybrid models are all fighting for dominance.
McKinsey’s 2026 service operations data highlights the economic engine behind this shift: an AI resolution costs $0.62, compared to $7.40 for a human agent. That gap is why Forrester reports that over 45% of organizations already use AI agents, with another 25% currently piloting them. IDC predicts that seat-based pricing will be largely obsolete by 2028 as the market pivots to performance-linked metrics – though companies should be wary that the efficiency gains they expect to keep for themselves are often transferred directly to the vendor through these new pricing structures.
Not everyone is sold on the math. A Forbes article by Chris Silver, ‘Outcome-Based Pricing: The Most Expensive Myth In Enterprise AI,’ argues that this model creates a classic principal-agent problem. Vendors are incentivized to maximize the billing metric – the number of resolutions – rather than the quality of the customer experience. This can lead to hidden permanent risk premiums and significant administrative overhead as companies struggle to reconcile complex, usage-based invoices.
The market is cluttered with noise. Gartner notes that of the thousands of vendors claiming agentic capabilities, only about 130 are genuinely agentic, with the rest engaging in ‘agent washing.’ That confusion contributes to a high failure rate: Gartner warns that over 40% of agentic AI projects will be canceled by the end of 2027 due to escalating costs, unclear business value, and inadequate risk controls. The promise of AI efficiency is often harder to realize than the marketing materials suggest.
The most direct consequence of this shift is human. When a company stops paying for a seat, the person who occupied that seat loses their place. Industry analysts estimate that AI agent deployments will eliminate the need for 20-35% of enterprise SaaS seats by the end of 2027. As companies move away from per-seat licensing, they are effectively decoupling their operational capacity from human headcount. The people who once performed these tasks are being replaced by automated workflows that are billed as a utility.
Vendors are throwing everything at the wall right now – outcome-based, consumption-based, per-agent subscription, and hybrid approaches are all in play. Sierra is opting for high-touch, custom contracts starting at $150,000 per year. Others are pushing self-serve, credit-based systems at pennies per action. There is no single winner, and the only certainty is that the old way of buying software is dying.
Gartner projects that by 2030, agentic AI could impact approximately 20% of enterprise SaaS spend, with 35% of point-product SaaS tools being replaced or absorbed by agents. The challenge for executives and operators is to navigate this transition without falling for the hype or getting locked into pricing models that favor the vendor. The cost savings are real – $0.62 versus $7.40 is not a rounding error – but so is the risk of outsourcing core business functions to a black-box billing model that charges by the outcome without guaranteeing one.