AI Agents: The New Workforce or Just Another Software Cost?

Microsoft's Rajesh Jha recently floated a bold idea at a tech conference. In the future, AI agents might be considered new employees, each requiring a software license. It's a concept that could flip the whole software-as-a-service (SaaS) pricing model on its head.

AI as Employees?

In a world where companies deploy fleets of AI agents, the notion that these entities could have their own identities, logins and inboxes, might not be so far-fetched. Jha posits that if AI agents are seen as individual users, software revenue could expand, not shrink. Imagine organizations with more AI agents than human employees, each agent effectively occupying a 'seat' that requires its own software license.

This isn't just a whimsical prediction. It's a potential strategic pivot in the ongoing SaaS pricing debates. Investors have been wary that AI could undercut seat-based pricing. If one human can manage an array of AI agents, why should companies pay for multiple licenses?

The Price of AI

Jha's scenario suggests a company with 20 employees might hold 20 Microsoft 365 licenses today. If each employee is assisted by five AI agents, and the workforce shrinks to 10 people, that could still mean 50 paid seats. The capex number is the real headline here, suggesting a different future for software economics.

Yet not everyone buys into this vision. Nenad Milicevic at AlixPartners offers a counterpoint. He argues that AI agents might reduce human interaction with software, thus decreasing the need for multiple licenses. If companies adopt this view, vendors could face pressure to adjust pricing models that no longer hold water.

The New Software Economics

Open platforms might become the real winners. They could charge for machine-based access without alienating customers. Companies unwilling to pay double for human and AI agent licenses might shift to competitors that allow AI to operate without added cost.

So, are AI agents just extensions of human workers, or should they be treated as autonomous entities? This debate could shape the next decade of software economics. The strategic bet is clearer than the street thinks. As AI continues to evolve, enterprise adoption strategies will need to adapt, fast.