Salesforce CEO Marc Benioff has repeatedly pushed back against warnings that AI could undermine the software-as-a-service (SaaS) business model. Now, as Salesforce with reporting accelerating demand for its AI products, the CEO is reaffirming his argument that rather than causing a “SaaSpocalypse,” the commoditization of AI models will increase the importance of enterprise applications and customer relationship management (CRM).

Benioff launched another forceful defense of enterprise software on the vendor’s second-quarter earnings call, arguing that the growing availability of AI models will strengthen the role of CRM rather than eliminate it.

Benioff pointed to growth in software seats, customer retention, contract lengths and AI usage as evidence that fears over AI agents destroying the traditional SaaS model have yet to materialize.

“This is the future of enterprise software, and I will tell you what it is not. This is not the SaaSpocalypse. We have been hearing about this for the last two quarters, these dire predictions about the end of software and how the models eat everything. But none of them have come true for us.”

The comments came as Salesforce reported second-quarter revenue of $11.35BN, up 11 percent year over year, while subscription and support revenue increased 12 percent to $10.82BN.

Annual recurring revenue for the Agentforce agentic AI platform reached $1.5BN, while Salesforce said customers generated 3.2BN agentic workspace units during the quarter, up 97 percent from the previous quarter.

AI Models May Be Commoditized, But CRM Remains the Context Layer

A central part of Benioff’s argument is that enterprises will increasingly consume AI through the software applications they already use rather than managing an expanding collection of foundation models themselves. The model will become increasingly abstracted from the customer.

“I believe what they are going to do is consume AI through packaged software… You are using Slack, you are consuming AI through packaged software. You are using Salesforce, you are consuming AI through packaged software.”

Benioff compared the position of AI models with other infrastructure that businesses typically do not interact with directly. “Do you even know what chip you’re using? Do you really know what data center you’re on? Do you know what router you’re on, what switch, what type of fiber, what cable you’re using, what interconnect you’re using?”

“At some level, we do get abstracted back to how the way the world used to be, which is we operate at a certain level of value. I believe the highest level of value is at the application layer,” Benioff added.

As the AI market becomes increasingly crowded with competing frontier and open models that have similar capabilities, the differentiation for enterprise buyers could increasingly sit in proprietary data, workflows, permissions, business rules, integrations and applications.

Salesforce is positioning CRM as a core part of that infrastructure.

“We are told the frontier models would eliminate the need for CRM,” Benioff said. “Instead, they depend on it, and we’re showing how 1 plus 1 can equal 3.”

Customer Vibe-Coding Experiment Reinforces Salesforce’s Argument

Salesforce used a customer’s experience with “vibe coding” to reinforce the point that AI-generated software will not automatically make enterprise CRM platforms obsolete.

David Friedberg, CEO of agricultural technology company Ohalo explained that his company had experimented with building its own CRM using AI coding tools before ultimately choosing Salesforce.

“You quickly realize just how much it takes to do maintenance, to do accounts, to do security. There is just much more to it,” Friedberg said.

The experience led Ohalo to distinguish between software it could build itself and foundational enterprise systems it would rather buy.

“We’re not going to vibe code CRM… We are going to build custom interfaces.”

Friedberg’s experience also offers an example of how AI coding tools can strengthen rather than displace an enterprise software platform. He said an employee named Ben used Claude and Cursor to customise Salesforce and get the company’s system running in less than a month.

Benioff seized on the example as evidence that AI coding tools could make Salesforce more valuable by reducing the effort required to configure and administer the platform.

“I think that has been the huge shock for us, that these new next-gen tools, the Cursors, the Replit, the Claudes of the world, make our product better,” Benioff said.

Friedberg, who credited Benioff as “a great marketing guy,” admitted to revising his view of the SaaSpocalypse argument.

“I was probably early on the SaaSpocalypse train… I always said I think SaaS was this temporary phenomenon between the founding of the Internet and the start of AI. But I think that it’s a little more nuanced than that.”

“I think there is still a SaaSpocalypse, but with a lowercase S rather than the uppercase S, for verticalized tools, where I think AI really allows you to rebuild something that is unique, which creates value for your business in a vertical,” Friedberg added.

The exchange gave Salesforce an example to support its broader argument that AI-generated applications do not necessarily eliminate the need for enterprise software. Instead, the company is positioning CRM as the governed foundation on which businesses can use AI to build increasingly customised workflows and interfaces.

Benioff’s response to the SaaSpocalypse narrative is built around Salesforce’s customer data. The company had been warned that software seats would decline as AI agents increasingly perform work on behalf of employees. Instead, seats across Agentforce Sales, Agentforce Service and Slack increased year over year.

“Apps are not dying. In fact, they are growing,” Benioff said.

Benioff highlighted near-record-low customer attrition, improving contract lengths and increased bookings for higher-end editions as evidence that customers continue to invest in Salesforce.

The company is also seeing substantially more activity from agents accessing Salesforce data. Agentic use through Model Context Protocol and CLI calls had increased sixfold, Benioff said.

That points to a potentially different model for enterprise software in which users may spend less time navigating individual applications while AI agents spend more time interacting with the underlying systems. Benioff argued that “it is a shift from software as the interface to software powering every interface.”

Data and Governance Become Part of the Value Proposition

Benioff argued that the growing capability of AI also makes the underlying enterprise data layer more important. “Now, we all know that if your data’s not right, your AI is not right.”

Salesforce’s platform contains hundreds of petabytes of enterprise context that can help agents understand a company’s customers, products and processes. The vendor is investing in data integration, harmonization and governance through its Data 360 platform and acquisitions including Informatica.

A model can generate an answer, but the enterprise application determines what data the model can access, which actions it can take, what permissions apply and how those actions fit into established business processes.

“You have to remember, models are probabilistic systems. They are non-deterministic systems,” Benioff said. “But our system, those four layers that we talked about with data and apps and semantics and agents, those are deterministic systems.”

“When you put those two things together, that is a level of value we have never seen before.”

Salesforce can argue there is no SaaSpocalypse while simultaneously adapting to a software model in which AI agents become an increasingly important source of application usage.

As agents carry out more work, the traditional link between software revenue and the number of human users becomes less relevant, with consumption, transactions and outcomes potentially taking a larger role in how enterprise software is priced. This may be the more significant development than the debate over whether the SaaSpocalypse exists.

Salesforce is defending the application layer while changing how that layer is used and monetized. The AI business is expanding much faster than the underlying software company, even as Salesforce continues to rely on its established application platform as the foundation.

Agentforce and Data 360 annual recurring revenue (ARR) reached nearly $3.9BN during the second quarter, growing more than 210 percent year on year, far outpacing the more modest 11 percent growth in overall revenue.

The company raised its revenue guidance for the 2027 financial year to between $46.1BN and $46.4BN, representing growth of roughly 10-11 percent from the previous financial year. The figures suggest AI is becoming a substantial growth engine for Salesforce, but it has yet to translate into a comparable acceleration in the company’s overall top-line growth.