As fears grow that artificial intelligence (AI) may kill the traditional software business model, a top executive at Microsoft has floated a new idea: make the AI agents pay for their own seats – just like in the case of human workers. For decades, the “Software as a Service” (SaaS) industry has relied on “seat-based pricing”, which is a system where a company pays a monthly fee for every human employee who uses a program.
‘Treat AI agents as human users’
With AI agents now capable of doing the work of multiple people, investors have panicked, fearing that smaller human headcounts will lead to shrinking profits for giants like Microsoft, Salesforce and Workday. However, Rajesh Jha, Microsoft’s vice president of experiences and devices, addressed these concerns at a recent conference by flipping the script. He suggested that in the near future, AI agents won’t just be tools; they will be digital employees with their own identities.Under Jha’s vision, an AI agent would have its own login credentials, email inbox, security permissions and access to specific software tools – making them like any other human user – and software companies may charge them on per agent basis. “All of those embodied agents are seat opportunities,” Jha explained, as per a report by Business Insider. He envisions a world where a company might have more AI agents than human workers – and every single one of them would require a paid license.
‘Protecting the SaaS business’
This ‘math’ behind this “Agentic AI” model may be a lifesaver for the tech industry. For example, if a company has 20 employees, it buys 20 licenses. If AI allows that company to shrink to 10 humans, software revenue would normally drop by half.However, if those 10 humans are managing 50 AI agents, and each agent requires a license, the software provider actually increases its sales from 20 seats to 60.If Jha’s vision takes hold, it could effectively wipe out “Claude fears” – a term loosely used for the anxiety that highly capable AI models like Anthropic’s Claude will make traditional software seats obsolete. The fear is: if AI makes each human worker vastly more productive, companies need fewer workers, and therefore fewer licences. Revenue shrinks. This fear came after Anthropic launched specialised plugins for use in judiciary. Jha’s argument is that this fear is based on a misreading of how AI will actually be deployed.