Dreamforce 2026 has officially wrapped, and this year’s event brought much more than a collection of new Salesforce features. Across three packed days, the conversation moved toward a different way of building, accessing, and managing business technology.
The biggest change was the growing connection between AI, business data, agents, applications, and the places where teams already work. Instead of treating these as separate pieces, Salesforce used Dreamforce 2026 to show how they can work together across the enterprise.
From the launch of AIforce and Koa to new Agentforce capabilities, Slack innovations, and deeper AI partnerships, Salesforce Dreamforce 26 gave businesses plenty to take away. So, let’s look at what happened across all three days and the developments that stood out.
Three Days That Reshaped the Agentic Enterprise
Day 1: AIforce Takes Center Stage
Day 1 established the main direction for Dreamforce 2026: Salesforce wants its data, business logic, workflows, and governance to be available wherever people and AI agents are working, rather than keeping everything inside one traditional interface.
AIforce was the headline announcement, introducing a live interface layer that connects Salesforce’s data, workflows, business logic, permissions, security, and governance with external AI experiences. This means users can interact with Salesforce capabilities without necessarily opening Salesforce itself.
- AI interfaces can reason across connected Salesforce information.
- Existing permissions and business rules continue to govern access.
- Users can take action instead of simply receiving information.
- MCPs, APIs, plug-ins, and skills can be used to create new experiences.
- Salesforce highlighted Zero Data Retention for business data used through AIforce.
Claudeforce expands the Salesforce-Anthropic relationship by bringing Salesforce context and actions into Claude. Salesforce in Claude includes 37 prebuilt sales skills designed around activities such as working with revenue context and updating pipelines.
- Salesforce data and workflows can be accessed through Claude.
- A prebuilt MCP server reduces the need for complicated custom integrations.
- Sales teams can work with CRM context directly through Claude.
- Additional integrations between Salesforce, Claude, and Slack are planned.
Koa brought another interesting layer to the AIforce story. Salesforce introduced it as its first CRM reasoning model, built specifically around the type of multi-step reasoning needed for CRM tasks.
Rather than treating CRM as another general knowledge domain, Koa is designed around business processes, tool usage, and complex tasks that agents may need to complete. Salesforce announced Koa alongside NVIDIA technology as part of its broader AI development.
- Built around CRM-specific reasoning.
- Designed for longer and more complex tasks.
- Supports the wider Agentforce architecture.
- Developed with NVIDIA technology.
“Models alone cannot run the enterprise. Models alone are not going to show us what’s possible. Models also are very probabilistic.”
— Marc Benioff, Chair and CEO, Salesforce
As AI moves deeper into business operations, governance cannot sit separately from the technology. AIforce was designed so requests run through existing permissions and business rules, giving organizations a familiar control structure as new interfaces are introduced.
- Existing Salesforce access controls continue to apply.
- Business actions route through Salesforce.
- Governance is built into the architecture.
- Zero Data Retention was highlighted as part of the AIforce model.
Day 2: Agents Start Doing More of the Work
“Hunter is cooking in the background. He’s going through all of your account information, all of those campaigns that your customers are engaging with, and he’s providing all of that rich context to your marketing and sales teams so they can move more early stage opportunities along.”
— Matthew Schultz, Senior Director of Product Marketing, Salesforce
- Works with account and engagement information.
- Supports outbound sales activities.
- Can continue working toward longer-term objectives.
- Gives teams additional context around early-stage opportunities.
Agent Optimizer focused on a part of the agent lifecycle that can easily be overlooked: what happens after deployment.
The capability brings observation, diagnosis, and improvement into the process, helping teams identify when agents are not performing as expected and make changes based on those results.
- Connects testing and monitoring.
- Helps identify underperforming agents.
- Supports ongoing optimization.
- Moves agent management toward a continuous process.
Headless capabilities showed how Salesforce data and business functionality can be exposed through APIs, MCP, and developer tools instead of being tied to one fixed user interface.
This gives developers more freedom to create experiences around existing Salesforce capabilities, while the underlying business logic and data remain connected to the platform.
- APIs and MCP can expose Salesforce capabilities.
- Agents can interact with business functions through different surfaces.
- Developers can create experiences outside the traditional interface.
- Business logic does not have to disappear simply because the interface changes.
Builder Central added another interesting piece to the development story. The idea is to make building applications and agents more accessible by allowing teams to describe what they need using natural language.
This could make the distance between a business requirement and a working solution much smaller, particularly for teams that do not want every new idea to begin with traditional development work.
Data 360 remained important throughout the Salesforce Dreamforce 2026 conversation because agents need more than isolated pieces of information. They need connected context to understand customers, business relationships, and what is happening across different systems.
- Connects business information.
- Helps provide richer context to agents.
- Supports more complete customer understanding.
- Gives AI more information to work with when handling complex tasks.
Salesforce also continued pushing AI into specific industries and business functions. The focus was not simply on creating a general AI assistant, but on applying agents and intelligence to particular workflows and requirements.
This approach appeared across areas such as sales, service, operations, life sciences, and other industry environments throughout the event. Salesforce’s official DF26 lineup also highlighted customer and industry-specific Agentforce use cases.
Day 3: Slack Becomes a Bigger Part of the AI Workplace
The final day brought many of the earlier ideas into the daily working environment. Slack became a major focus, showing how conversations, data, agents, coding, and business actions can increasingly happen together.
Slackforce Surfaces allows users to describe what they want and generate interactive dashboards, reports, and other experiences using Salesforce and connected information. Instead of producing only a text answer, the experience can become something teams actually work with.
- Brings Salesforce information into Slack.
- Uses natural-language prompts.
- Supports interactive dashboards and reports.
- Lets teams explore and work with the information.
- Keeps the experience connected to the conversation.
Slack Code extends the same idea into engineering. AI coding agents can work inside dedicated Slack channels, allowing developers and teams to keep conversations, code-related work, and AI activity together instead of scattering them across multiple tools.
- Creates a shared environment for developers and AI agents.
- Supports agents such as Claude Code, Devin, GitHub Copilot, ChatGPT, and Vercel agents.
- Makes AI-assisted development more collaborative.
- Keeps project discussions connected to the development process.
“Slack code democratizes building… Coding is accessible to everyone now.”
— Katie Stigman, VP, Product Management, Slack, Salesforce
With businesses expected to use more agents across different functions, managing those agents becomes increasingly important. Agent Fabric focuses on areas such as discovering, connecting, testing, governing, and monitoring agents.
- Helps organizations understand their agent ecosystem.
- Supports governance and monitoring.
- Makes it easier to connect agents.
- Brings more visibility into how agents are being used.
Marketing was another major part of Day 3, with Salesforce showing how agents can support activities beyond content creation. Campaign optimization, inbound conversations, customer engagement, and AI-search visibility were all part of the wider discussion.
The direction is moving toward marketing agents that continue working after a campaign or interaction begins, rather than simply generating something and stopping there.
“It works end to end in your flow of work. It creates onbrand channel ready content in a new agentic editing experience and then assembles your campaign choosing the right timing message and channel for every customer. Then it keeps working after launch.”
— Natalie Matthysse, Product Marketing Director, Salesforce
Palmata highlighted another emerging marketing requirement: understanding how brands appear in AI-generated answers.
The tool focuses on measuring brand mentions, citations, competitors, and AI perception, helping marketers understand how their content and brand are represented as search increasingly moves toward AI-generated responses.
“Palmada is our latest innovation that helps you understand, measure, and improve your answer engine optimization so that you can truly identify what is going to work and help improve how you show up in searches, giving your marketers clear recommendations and guidance”
— Kimberly Gong, Vice President, Marketing, Contentful
Dreamforce 2026: Shift From AI Tools to an AI-Driven Future
Dreamforce 2026 showed that AI is moving beyond being a tool employees use occasionally. It is becoming part of the workflow itself, understanding context, working across systems, and increasingly taking responsibility for tasks that once required constant human involvement.
The next stage could be even more transformative. Instead of asking which software to open, teams may simply describe what they want to achieve while AI connects the right data, agents, workflows, and applications behind the scenes to make it happen.
And this is where the future gets interesting. As AI becomes more connected and capable, business software may become less visible while intelligent work becomes more seamless. Salesforce Dreamforce gave us a glimpse of that future; what comes next could redefine how work gets done.
