“Do AI Employees Need Salaries?” Microsoft Ignites the ‘Paywall Debate’

The burgeoning field of artificial intelligence is at a crossroads, and Microsoft's recent musings signal a potential seismic shift in how we pay for AI services. As AI agents become increasingly capable of handling complex, multi-step tasks, the question arises: should they be treated as 'employees' deserving of compensation? Microsoft's Rajesh Jha, VP of Experiences and Devices, has ignited this debate by suggesting that AI agents could become a new basis for paid accounts, moving beyond the traditional per-user model.

“

Is it necessary to pay AI employees salaries?

This proposed 'per AI agent' or 'per seat' model, as floated by Jha at the Microsoft Build conference, stands in stark contrast to the current industry standard. Today, most AI companies operate on a per-account subscription basis, where a single account grants access to the service, regardless of how many AI instances or agents a user might deploy within that account. Microsoft's potential pivot suggests a future where the sheer number of AI agents a business utilizes could directly impact its software licensing costs, potentially leading to a significant increase in expenses for heavy users.

“

There is a possibility that all AI agents will soon become opportunities for paid accounts.

However, this innovative pricing strategy is not without its critics. Some industry analysts express skepticism, arguing that charging for each AI agent could ultimately stifle adoption. The core argument is that the very efficiency of AI agents—the ability for one manager to oversee several agents to accomplish work previously done by many humans—will lead businesses to drastically reduce their overall license count to save costs. If Microsoft and others begin charging per agent, businesses might simply opt for fewer agents or seek out competitors who maintain a per-user or per-account model, potentially leading to a decline in revenue for the AI providers.

“

AI agents will eventually reduce the number of users, leading to a sharp decline in license demand.

This debate underscores the transitional phase the AI industry is navigating. As highlighted by insights from organizations like AlixPartners and communities like SaaStr, the fundamental business models for AI are being re-evaluated. The discussion is moving beyond simple metrics like user counts or active agents towards more nuanced approaches, such as performance-based billing. While performance-based models offer a direct link between value delivered and cost incurred, they also introduce revenue uncertainty for businesses. Ultimately, the industry is grappling with defining what exactly constitutes 'value' in the age of AI and how to monetize it sustainably, a challenge that requires a fundamental redesign of existing business frameworks.

“

We are in a transitional phase towards AI agents becoming a complete product.