While investors fret over a potential software rout due to ‘explosion’ of AI technology, one of Microsoft’s top executive has pushed back against fears of an enterprise software decline. According to Rajesh Jha, executive vice president of Microsoft’s Experiences + Devices Group, AI agents will still require software licenses, which addresses concerns that AI-driven job cuts would destroy seat-based revenue. Jha argued that as companies deploy thousands of digital agents, the number of ‘paying users’ could actually increase, even as traditional human headcounts fall.
Rajesh Jha’s logic behind the argument
The key insight is straightforward: In the future, Jha envisions, AI agents will be independent actors operating inside business software systems, complete with their own logins, inboxes and digital identities. And if they function as users, they may need to be licensed as users.“All of those embodied agents are seat opportunities,” Jha said, using the industry term for paid software licences. He explained it with an example: A company today with 20 employees might buy 20 Microsoft 365 licences. Now imagine that company deploys five AI agents for each worker and, as a result of that increased productivity, cuts its human workforce to 10 people. Under Jha’s model, that company could still be paying for 50 seats, which is the 10 remaining humans plus the forty agents working alongside them. In other words, a company that lays off half its workforce due to AI might end up spending more on software, not less because the AI agents filling the gap each require their own place in the system.
Why this matters for the software industry
The argument arrives at a moment of anxiety across the enterprise software sector as investors have been questioning whether the rise of AI poses an existential threat to seat-based pricing, the very model by which software companies charge per user, per month or per year. The fear is: if AI makes each human worker vastly more productive, companies need fewer workers, and therefore fewer licences. Revenue shrinks. The business model that has made enterprise software one of the most profitable sectors in technology begins to crack.Jha’s argument is that this fear is based on a misreading of how AI will actually be deployed. The assumption that AI reduces the number of software users is only correct if you define users as humans.
