The week‑long conference capped off with a high‑energy Dreamfest concert at Oracle Park, where Usher and Gwen Stefani performed for a packed crowd, underscoring Salesforce’s penchant for blending tech showcases with pop‑culture spectacles.
The headline announcement came from co‑founder Parker Harris, who introduced AIforce as the next step in the company’s “agentic AI” strategy. The service is described as a live interface that enables AI agents – from Claude to OpenAI models – to query Salesforce data, trigger workflows and apply business logic through APIs, command‑line tools and a model‑context protocol, eliminating the need for a traditional web or app UI.
AIforce builds on Salesforce’s earlier Headless 360 offering, which already exposed databases to agents via APIs and CLIs, and follows the 2024 launch of Agentforce and the recent rollout of Claudeforce earlier this month. Harris summed up the vision: “Why should you ever log into Salesforce again? Maybe you never will.”
Trust was a recurring theme throughout the keynote, with executives acknowledging that giving autonomous agents direct access to mission‑critical data raises security and governance questions the company says it is addressing.
The move arrives amid growing analyst warnings that agentic AI could upend the SaaS model. Gartner research cited at the event projected up to $234 billion of application spending could be exposed to “agentic arbitrage” by 2030, a scenario where AI agents bypass legacy dashboards entirely.
Salesforce’s CEO Marc Benioff dismissed talk of a looming “SaaSpocalypse,” but the company’s stock had already taken a hit earlier in 2026, sliding 40 percent after Anthropic unveiled Claude Cowork, a competitor that threatened to erode traditional software revenue streams.
Industry observers see AIforce as Salesforce’s attempt to stay ahead of the curve, positioning the platform as the backbone for AI‑driven workflows rather than a conventional CRM interface. If successful, the shift could reshape how enterprises interact with their data, but it also raises questions about data privacy, model bias and the reliability of autonomous agents in critical business processes.