As software architecture undergoes its most radical transformation since the dawn of the cloud, the industry is grappling with a fundamental shift in how value is captured. Vijay Raina, a seasoned specialist in enterprise SaaS technology and software design, sits at the intersection of this evolution, helping organizations navigate the transition from human-centric tools to autonomous systems. With a background rooted in building scalable architectures, he understands that the current friction isn’t just about the technology—it’s about a legacy billing framework that was never meant to handle the speed of AI. In this discussion, we explore the collapse of the per-seat model, the rise of “Work-as-a-Service,” and the massive financial infrastructure gap that threatens to leave even the most innovative vendors behind.
How does the traditional per-seat pricing model create a counterintuitive trap for vendors who successfully deliver high-value, efficient AI?
The per-seat model was built on a simple, comforting logic: one human employee equals one license, usually ranging from $20 to $50 a month. But in our current era, where AI is capable of handling entire workflows autonomously, this model turns a vendor’s success into a financial penalty. Imagine a B2B platform that manages customer support and successfully automates 80% of all incoming tickets; that is an incredible win for the client, who can then reduce a team of 50 reps down to just 10. For the software provider, however, this extraordinary ROI results in an 80% drop in subscription revenue because there are fewer “seats” to charge for. It creates a sickening feeling for a CEO to realize that the more efficient their product becomes, the less money they make. We are essentially watching a system where vendors are incentivized to keep human labor high, which is the exact opposite of what the market demands right now.
In what ways do autonomous AI agents fundamentally disrupt the concept of a “user” compared to the human employees these systems are replacing?
We have to move past the idea of the user as someone who logs into a dashboard from 9 to 5 to check boxes or generate short summaries. Autonomous agents are entirely different beasts; they don’t get tired, they don’t take breaks, and they execute hundreds or even thousands of actions every single minute without any human clicking a button. These agents are reading contracts, executing complex API calls, and resolving support tickets end-to-end at a pace that no human could ever match. Because they don’t consume software at a human cadence, trying to slap a “seat” price on them is like trying to charge a high-speed trading algorithm the same way you’d charge a person using a calculator. This discrepancy creates a massive margin drain for vendors who might charge a flat $50 monthly fee for an agent that is actually consuming $100,000 worth of operational value and massive LLM token overhead behind the scenes.
What are the primary differences between consumption-based models and the emerging “Work-as-a-Service” paradigm?
The industry is splitting into two paths to solve this monetization crisis, both of which move away from static monthly recurring revenue. The first is usage-based pricing, where the cost is tied directly to technical metrics like tokens consumed, data volume processed, or API executions to ensure the vendor’s revenue scales with their infrastructure costs. The second, more radical path is “Work-as-a-Service,” where the software is priced exactly like an employee based on finished outcomes. Instead of paying for access to a tool, the customer is billed per qualified sales lead, per processed invoice, or per resolved ticket. This shift is profound because it demands that the software actually perform the task rather than just assist a human in doing it, aligning the price directly with the economic value generated for the business.
Why is the transition to these new models described as an “operational nightmare” for existing SaaS companies?
Most companies are currently operating on legacy billing engines that were designed for the simple, predictable world of flat-rate subscriptions. These systems are completely ill-equipped to track micro-transactions or calculate dynamic usage against complex enterprise contracts in real time. It is a massive technical debt problem; these engines can’t handle the pressure of passing through unpredictable token costs or executing the metering required for “Work-as-a-Service.” On the buyer side, this often leads to “bill shock” for CFOs who are used to fixed costs and suddenly see a spike in invoices because their AI agent was particularly busy that month. We’re seeing companies like Vayu raise $7 million in seed funding specifically to solve this because the hardest part of the AI era isn’t just building the agent—it’s building the financial pipes to make sure the vendor isn’t leaking revenue.
How will the ability to master dynamic monetization become a greater competitive advantage than the actual AI technology itself?
We are witnessing the end of the steady-state SaaS era where a great product was enough to guarantee growth. The winners of the next decade won’t necessarily be the ones with the flashiest AI demos or the largest language models; they will be the companies that figure out how to align their pricing with the real economic value they create. If you are still clinging to per-seat subscriptions while your AI is doing 90% of the labor, you will watch your revenue crumble under the weight of your own automation. Mastery of agentic pricing allows a vendor to unlock unprecedented margins because they are capturing a piece of the labor cost they are replacing rather than just a utility fee. Monetization can no longer be an afterthought—it is now the core architectural challenge for anyone building enterprise software.
What is your forecast for the evolution of SaaS monetization through 2028?
By 2028, I expect the “per-seat” model to be a relic of the past for any software that incorporates autonomous agents. We will see a standardized “hybrid” model become the norm, where a base platform fee covers basic access, but the majority of revenue is generated through high-margin outcome-based milestones. The CFO’s office will transition from managing licenses to managing “digital labor budgets,” treating AI agents as a variable workforce rather than a fixed software expense. Companies that fail to implement sophisticated metering and real-time billing by then will likely face insolvency as their infrastructure costs for running AI outpace the stagnant revenue from their old-school subscription tiers. The shift from selling tools to selling results is inevitable, and the infrastructure to support that shift will be the most valuable part of the tech stack.
