How can SaaS companies survive in the post-AI era? Microsoft executives aren’t worried: even AI agents have to buy licenses

The AI wave is disrupting the global enterprise software market, and the stock prices of giants like Salesforce, Workday, and others have dropped significantly this year so far. The alarm bells for “SaaSpocalypse (SaaSpocalypse)” are ringing loudly. However, a very different perspective has recently been proposed by Microsoft executive Rajesh Jha: the rise of AI agents may bring brand-new growth opportunities for SaaS.

The AI wave hits the SaaS industry, and the stock market reflects panic first

In recent months, the term “SaaSpocalypse” has spread rapidly in the tech world, referring to a crisis in which AI could end the traditional enterprise software as a service (SaaS) business model. At the time, when Anthropic launched the Claude Cowork AI Agent—highlighting its advantage in autonomously handling complex tasks—it prompted outside questioning about whether SaaS subscription-based models will live or die.

As Vibe coding technology becomes increasingly popular, people without a programming background can now build an application in just a few minutes using AI tools. If a company can build software to manage sales or HR processes based on its own needs, what reason is there to spend hundreds of thousands of dollars to buy subscription licenses from Microsoft (Microsoft)?

(Anthropic’s new Claude Cowork triggers selling pressure, software and financial services stocks evaporate $285 billion in a single day)

A new seat-billing business opportunity may be emerging? Microsoft executives propose a “pay-per-AI-agent” model

Just as many people were predicting the end of the SaaS era, Rajesh Jha, head of Microsoft’s Experiences and Devices division, presented a completely different view in a recent meeting. He believes that as enterprises begin large-scale deployments of AI agents, each agent may require its own identity authentication:

Including system login accounts, email addresses, and even paid usage licenses on software platforms (what the industry calls “seats (seat)” ). All these concrete agents are seat-billing opportunities.

He further pointed out that a company with 20 employees today buys 20 Microsoft 365 licenses; but if each employee is assigned 5 AI agents in the future, even if the total headcount is reduced to 10, the number of paid seats could still remain at 50. Therefore, AI will not compress the software licensing market size; instead, it may expand it significantly.

Will AI agents expand the market or crush pricing?

However, not everyone buys into this theory. Nenad Milicevic, a partner at management consulting firm AlixPartners, believes that the widespread adoption of AI agents will drastically reduce how often humans actually operate software, and the number of licensed seats will shrink significantly. In the future, an enterprise’s operating model may only need one employee to supervise a small number of AI agents, rather than the past scenario where 20 employees log into different systems each on their own.

If software vendors try to charge extra for machine-access services, they may instead push customers toward competing open platforms that allow AI agents to access freely without additional barriers. For customers, that increases their bargaining leverage.

Market research firm IDC predicts that the pure seat-based pricing model will be a thing of the past before 2028, replaced by a consumption-based pricing model that charges based on usage or actual outcomes. This includes Salesforce and Microsoft, which are both carefully studying how to guide customers to transition to the new pricing model.

Where is the root of the problem: Are AI agents “employees”?

The core of the industry debate can be summarized into a simple question: can AI agents be considered “enterprise employees”? This question has not yet been settled, but the answer will directly determine the economic model and competitive landscape for the global enterprise software industry over the next decade.

Workday CEO Bhusri said in a earnings call that the two AI companies—Anthropic and OpenAI—are widely seen as the biggest threats to SaaS, and they themselves are also paid customers of Workday. This detail is also the most intriguing footnote in this survival battle.

How can SaaS companies survive in a post-AI era? Microsoft executives aren’t worried: AI agents also need licenses First appeared on Chain News ABMedia.