Salesforce plans to introduce a proprietary dynamic AI-powered interface within the next 12 months, CEO Marc Benioff said at The Six Five Summit in San Francisco on Tuesday. The move reflects the company’s push to transition enterprise software from deterministic systems to AI-driven workflows that adapt to user needs in real time.
Benioff outlined three phases of enterprise AI development: AI 1.0, which relies on prompt-based models using public data; AI 2.0, which introduces agentic systems capable of executing tasks autonomously; and AI 3.0, which integrates these agents with dynamic interfaces that merge structured enterprise software with probabilistic AI. He emphasized that Salesforce is currently operating in AI 2.0, with plans to advance to AI 3.0 within a year.
The company’s existing AI agents—including Casey for customer service, Slackbot for Slack users, and Marshall for supply chain operations—demonstrate the practical applications of this strategy. Casey has resolved 5 million customer service issues on Salesforce’s help platform, while Slackbot supports tens of millions of users. Dell’s Regrello system, which interacts with 20,000 suppliers daily, further illustrates the scalability of agentic AI in enterprise environments.
Salesforce’s financial performance underscores its capacity to invest in AI development. The company reported nearly $43 billion in revenue over the trailing twelve months and a market capitalization of $169 billion. Gross profit margins stand at 77.6%, revenue growth is 11%, the PEG ratio is 0.61, and free cash flow yield is 9%. The stock closed at $205.69 on Tuesday, down 1.61%.
Benioff also highlighted the importance of governance in AI deployment, noting that Salesforce has hired thousands of new graduates to build controlled, secure AI systems. He contrasted unstructured “vibe coding” with development on deterministic enterprise platforms, arguing that the latter ensures security and compliance. “You can vibe code all by yourself with no controls and no security, or you can vibe code on top of one of these deterministic systems,” he said.
The executive criticized market narratives that overstate AI readiness, stating, “Everybody’s talking their book.” He described a common enterprise scenario where companies juggle multiple software ecosystems—Salesforce, Microsoft, SAP, Teams, Slack, and hyperscalers—without having initiated meaningful AI transformation. “Hey, I’ve got my Salesforce, I’ve got Microsoft, SAP, maybe I’ve got my data lake,” he said. “I’m using Teams or I’m using Slack and a hyperscaler, and I still have not really started my AI transformation.”
