Dreamforce 2026 Recap | San Francisco | Salesforce, AIforce, Agentforce, Claudeforce, and Enterprise AI
Held September 15–17, 2026 at the Moscone Center in San Francisco, Dreamforce 2026 signaled the start of a new era in Salesforce’s AI strategy.
This year, the stage wasn’t just about Salesforce’s new products. Salesforce CEO Marc Benioff sat down with some of the most prominent names in the AI world (OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and NVIDIA CEO Jensen Huang) to discuss the future of AI, enterprise adoption, and safety debates.
Dreamforce 2026’s shared message was clear: AI is no longer just an assistant that generates answers, it’s becoming a layer that understands company data and business processes, takes action, and runs the work itself.
“AI is creating an interface revolution.” — Marc Benioff
1. Salesforce’s big new move: AIforce
One of the most important announcements at Dreamforce 2026 was AIforce. Salesforce is positioning AIforce as a new “live interface layer” that carries the data, workflows, business rules, permissions, and security architecture companies have built inside Salesforce over the years into different AI interfaces. The shift here is significant. Until now, users would log into a CRM screen, find the relevant record, and take action. In Salesforce’s new vision, the interface is becoming increasingly dynamic. The user or AI agent states what they want to do, and the system assembles the necessary data, process, and action on the spot.
In other words: the CRM is no longer just an app you go to, it’s becoming an intelligence layer that runs wherever the work happens.
2. Agentic Enterprise is no longer just a vision — it’s becoming an architecture
Salesforce has been using the term “Agentic Enterprise” for the past few years. At Dreamforce 2026, this approach turned into a more concrete architecture. At the center of this structure are Data 360, Customer 360, Agentforce, and AIforce. Data 360 enables agents to understand company data and customer context, while Customer 360 brings business rules, processes, permissions, and actions into the system. Agentforce represents the digital workforce layer that carries out the work, while AIforce represents a dynamic experience that lets different AI interfaces make use of this entire structure.
The core shift here is that AI is no longer a separate tool working alongside Salesforce. AI is increasingly embedding itself directly into the company’s data and operations layer.
3. OpenAI and Anthropic were on the same Dreamforce stage
One of the notable aspects of Dreamforce 2026 was that the CEOs of two major AI companies appeared on the same event program. OpenAI CEO Sam Altman spoke with Marc Benioff about where AI currently stands, next-generation models, and how this intelligence can turn into reliable tools for businesses.
Anthropic CEO Dario Amodei, for his part, discussed the Salesforce–Anthropic partnership and the Claudeforce vision with Benioff.
There’s an important strategic message here for Salesforce: rather than building an ecosystem tied to a single AI model, the company is creating a more open structure in which different models can work on top of Salesforce’s trusted data and business processes.
4. Claudeforce, Koa, and the new enterprise AI layer
Dreamforce 2026 wasn’t just about Agentforce announcements. One of the new technology layers Salesforce highlighted was Claudeforce. Through the Salesforce–Anthropic partnership, Claude’s advanced reasoning capabilities will be able to work more directly with Salesforce data and business processes.
Another major announcement was Koa. Developed through a Salesforce–NVIDIA partnership, Koa was introduced as the first CRM reasoning model Salesforce has built specifically to reason over CRM data. Alongside this, Salesforce also spotlighted its Trusted Enterprise AI Harness approach, aimed at helping companies manage their different AI agents in a controlled way.
In fact, all of these announcements respond to the same problem: having more powerful AI models isn’t enough on its own. Models need to operate within the right data, the right context, security, and governance.
5. Data is still AI’s most critical issue
Even as product names changed at Dreamforce 2026, one fact stayed the same: the quality of AI still depends on the quality of the data feeding it. That’s why Salesforce keeps Data 360 and the Zero Copy approach at the center of its agentic architecture.
With Zero Copy, companies can use data that lives in different systems for AI processes without having to constantly move it into Salesforce.
This approach matters especially for large companies, since a significant portion of enterprise data sits outside Salesforce. For AI agents to make the right decisions, they need not just the model itself, but a reliable context that captures the company’s customers, processes, and business rules.
What was the real message of Dreamforce 2026?
Looking at the overall picture coming out of Dreamforce 2026, we can see that the core question in the AI world is starting to shift. In the early period, the question was: “What can AI tell us?” Then it became: “What tasks can AI perform?”
Now, the question companies need to answer is more strategic: “Which processes actually create value when handed off to AI?”
The structure Salesforce is trying to build with Agentforce, Data 360, AIforce, Claudeforce, and Koa is focused exactly on this point.
The Inspark Perspective
From our point of view, the most important message of Dreamforce 2026 is that AI can no longer be treated as a standalone technology project.
For an agent to run a process end-to-end, that process first needs to be clearly defined, fed with the right data, and mapped out at which point it requires human oversight. That’s why, going forward, success won’t be measured by “how many agents we’re using,” but by which processes we’re making faster, more accurate, and more measurable.
Salesforce’s message through Dreamforce 2026 is quite clear on this front: AI is no longer just a new feature added inside Salesforce; it’s becoming a layer that redefines how Salesforce and enterprise software as a whole will work.
Sources
