For years, the bedrock of the enterprise software industry has been the “seat.” Whether it is a productivity suite, a CRM, or a project management tool, the business model is almost universally the same: charge a monthly or annual fee per human user. It is a predictable, scalable engine that has fueled the meteoric rise of the SaaS (Software as a Service) era.
But as generative AI begins to automate complex workflows, a quiet panic has settled over Wall Street. The fear is simple: if AI makes a single employee as productive as five, companies will hire fewer people. Fewer people mean fewer seats, and fewer seats mean a collapse in revenue for the very software companies building the AI. This tension has turned the conversation around AI agents and software licensing into a pivotal debate over the future of the digital economy.
Rajesh Jha, Executive Vice President of Microsoft’s Experiences + Devices Group, is pushing back against this narrative. Jha argues that AI will not kill the seat-based model; instead, it will expand it by redefining what a “user” actually is.
Redefining the ‘User’ in the AI Era
The core of Jha’s argument rests on the transition from AI as a “tool” to AI as an “agent.” A tool, like a basic chatbot, is something a human uses to receive a task done faster. An agent, however, is an autonomous entity capable of executing multi-step goals, managing its own schedule, and interacting with other software systems independently.
From a technical perspective—a nuance often overlooked in investor calls—an autonomous agent cannot simply exist as a feature within a human’s account. To operate securely and effectively within a corporate environment, an agent requires its own digital identity. It needs its own login credentials, its own inbox to receive communications, and its own set of permissions to access specific databases.
“All of those embodied agents are seat opportunities,” Jha said, using the industry shorthand for paid licenses. By treating these agents as independent actors, software companies can shift their billing from “human users” to “digital users.”
The Math of the Digital Workforce
To illustrate this shift, Jha provides a scenario that flips the traditional fear of AI-driven layoffs on its head. In a conventional setup, a company with 20 employees would purchase 20 licenses for a suite like Microsoft 365.
Under an AI-integrated model, that same company might deploy five specialized AI agents for every human worker to handle repetitive tasks, data analysis, and scheduling. This surge in productivity might allow the company to reduce its human headcount to just 10 people. On the surface, this looks like a 50% loss in licensing revenue for the software provider.
However, if each of the 40 AI agents is treated as a licensed “seat,” the company is now paying for 50 seats—10 humans and 40 agents. In this model, the software provider actually sees a 150% increase in license volume, even as the human workforce shrinks.
| Metric | Traditional Model | Agent-Based Model |
|---|---|---|
| Human Workforce | 20 Employees | 10 Employees |
| AI Agents Deployed | 0 | 40 Agents |
| Total Paid Seats | 20 | 50 |
| Revenue Impact | Baseline | Significant Increase |
An Existential Pivot for Enterprise Software
This perspective is a direct challenge to the “existential threat” narrative currently circulating in the tech sector. For years, the Gartner-style projections of software growth have been tied to the growth of the global workforce. If the workforce plateaus or declines due to automation, the growth ceiling for SaaS companies theoretically drops.
Jha’s logic suggests that the ceiling isn’t dropping; it is being replaced by a much higher one. The “user” is no longer a biological constraint but a functional one. If a business needs 1,000 autonomous processes to run its supply chain, it will need 1,000 digital identities, regardless of how many humans are overseeing them.
This shift, however, creates new complexities for the industry:
- Pricing Friction: Companies may resist paying the same “per-seat” price for an agent as they do for a human, leading to a tiered pricing structure.
- Security and Governance: Managing thousands of autonomous digital identities introduces significant cybersecurity risks, requiring more robust identity and access management (IAM) tools.
- Value Realization: Software vendors must prove that these agents provide enough tangible value to justify the cost of additional licenses.
Who is affected by this shift?
The primary stakeholders in this transition are the Chief Financial Officers (CFOs) and Chief Information Officers (CIOs). For the CFO, the trade-off is clear: replacing expensive human payroll with relatively cheaper software licenses. For the CIO, the challenge is the infrastructure—ensuring that these “embodied agents” can communicate across different platforms without creating security loopholes.
For the employees, the implication is more sobering. Even as Jha’s model is a win for software revenue, it validates the idea that AI can drastically reduce the number of humans required to perform a given volume of work.
Disclaimer: This article discusses business models and investment trends in the technology sector. It is provided for informational purposes only and does not constitute financial or investment advice.
The industry is now watching closely as Microsoft and its competitors begin to integrate more “agentic” capabilities into their platforms. The next major indicator of this trend will be the upcoming pricing updates for Copilot and the potential introduction of specific “Agent Licenses” in future enterprise agreements.
Do you think AI agents should be billed as “seats,” or is this just a way for software companies to maintain their margins? Let us know in the comments.
