Dreamforce 2026 | Salesforce CEO Benioff unveils ‘AIForce’ as protesters demand end to ICE contracts

 

Salesforce’s annual Dreamforce conference in San Francisco has once again become a major stage for enterprise artificial intelligence. At Dreamforce 2026, Salesforce CEO Marc Benioff unveiled AIForce, a new AI-powered interface layer designed to bring Salesforce’s business data, workflows and AI agents into different interfaces and applications.

However, the technology announcements were accompanied by controversy outside the conference venue. Protesters gathered near the Moscone Center demanding that Salesforce end contracts connected to U.S. Immigration and Customs Enforcement (ICE). At the same time, AI safety became another major topic at the event, with technology leaders debating how quickly artificial intelligence should advance and how its risks should be controlled.

Dreamforce 2026 is taking place from September 15 to September 17 in San Francisco, bringing together Salesforce customers, developers, technology leaders and AI companies.

Salesforce Unveils AIForce

The biggest product announcement from the opening day was AIForce.

Benioff described AIForce as part of what he called an “interface revolution.” Rather than forcing users to work through a traditional fixed Salesforce interface, AIForce is designed to allow people and AI agents to interact with Salesforce capabilities wherever they work.

According to Salesforce, AIForce can connect an AI interface with the company’s existing:

• Business data

• Workflows

• Business logic

• Permissions

• Security controls

• Governance systems

• AI agents

The idea is to make Salesforce’s enterprise capabilities available through different AI interfaces instead of keeping them restricted to a conventional CRM screen.

This is particularly important as businesses increasingly use conversational AI and autonomous agents to perform tasks that previously required employees to navigate multiple software applications.

How AIForce Works

AIForce is designed as an additional layer on top of Salesforce’s existing AI architecture.

Salesforce says its broader “Agentic Enterprise” architecture includes Data 360, Customer 360 and Agentforce, with AIForce acting as a new interface layer.

Data 360 provides unified business information and context, while Customer 360 provides business processes, permissions and logic. Agentforce provides AI agents capable of carrying out business tasks.

AIForce is intended to bring these capabilities into different AI interfaces.

For example, instead of opening a CRM application, searching through customer records and manually checking workflows, a worker could potentially ask an AI interface for the required information and have the system reason across Salesforce data and business rules.

Salesforce says AIForce is designed to be:

Intelligent and dynamic: AI agents can reason across data, business logic and conversations.

Secure and governed: Existing permissions and business rules continue to control what users and agents can access.

Composable: Developers and builders can create interfaces and agents using APIs, MCPs, plug-ins and other tools.

Zero Data Retention: Salesforce says business data used to answer a request is not retained by the model provider.

ClaudeForce and SlackForce

AIForce is not being introduced as an isolated Salesforce product.

Salesforce says its initial rollout includes Claudeforce, Slackforce and Agentforce Coworker.

Claudeforce is particularly significant because it brings Salesforce capabilities into Anthropic’s Claude ecosystem. The strategy reflects a broader change in enterprise software: companies increasingly want employees to access business information through AI assistants rather than traditional application interfaces.

Salesforce is therefore attempting to make its CRM platform useful even when users are interacting with another AI system.

The company has also been expanding its relationships with major AI and cloud companies, making interoperability an increasingly important part of its enterprise AI strategy.

Salesforce’s Bigger Shift Toward Agentic AI

AIForce is part of Salesforce’s broader attempt to transform its business around agentic AI.

Traditional software generally waits for a human user to click buttons, enter information and select actions. Agentic software aims to allow AI systems to understand a goal, reason about available information and perform multiple steps.

Salesforce has positioned Agentforce as its AI workforce, while AIForce focuses more heavily on how people and agents interact with Salesforce's underlying data and business capabilities.

The company believes this could change how enterprise software is used.

Instead of learning dozens of menus and dashboards, employees could increasingly communicate with software using natural language.

For businesses, the potential benefit is faster access to information and automation of repetitive processes. However, it also increases the importance of security, permissions and monitoring because AI agents may be able to take actions on behalf of employees.

Protesters Demand Salesforce End ICE Contracts

While Salesforce was presenting its AI strategy inside the Moscone Center, protesters gathered outside the venue.

The demonstrations focused on Salesforce’s relationships with U.S. government agencies, particularly contracts associated with U.S. Immigration and Customs Enforcement (ICE).

Reports from the event said protesters called on Salesforce to end contracts connected to ICE and raised broader concerns about corporate responsibility and government technology contracts.

The protest added a political and social dimension to an event otherwise dominated by AI announcements.

The issue is not entirely new. Salesforce has faced criticism from activists over its government-related technology contracts, while the company has continued to emphasize its role as an enterprise software provider.

The demonstrations show that the debate surrounding AI and technology companies extends beyond product capabilities. Questions about how enterprise technology is used by governments and public institutions are increasingly becoming part of the technology industry's public discussion.

AI Safety Debate Takes Center Stage

Another major theme at Dreamforce 2026 was the growing debate over AI safety.

Anthropic CEO Dario Amodei appeared at the event and argued for greater caution around the speed of AI development. His position comes amid increasing concern among some AI researchers and executives that increasingly capable AI systems could create risks that are difficult to control.

The discussion reflects a wider disagreement within the technology industry.

Some AI leaders have called for stronger safeguards, independent evaluations and greater coordination around AI safety. Others, including Nvidia CEO Jensen Huang, have argued that companies can continue advancing AI while addressing safety through engineering practices rather than broad restrictions.

At Dreamforce, these different perspectives highlighted one of the central questions facing the AI industry: How quickly should increasingly powerful AI systems be developed and deployed?

Jensen Huang and Enterprise AI

Nvidia CEO Jensen Huang also featured prominently in Dreamforce discussions.

Nvidia supplies much of the computing hardware that powers modern AI systems, making Huang an important voice in the debate over AI development.

At Dreamforce, Huang argued against slowing AI progress and emphasized the possibility of combining rapid innovation with responsible engineering and product safety.

His position contrasts with the more cautious approach advocated by Amodei.

This difference is significant because Salesforce, Anthropic and Nvidia occupy different but interconnected positions in the AI ecosystem. Salesforce is focused on enterprise applications, Anthropic on foundation models, and Nvidia on the computing infrastructure used to train and run those models.

What Dreamforce 2026 Means for Enterprise AI

The announcements at Dreamforce show that enterprise AI is moving beyond simple chatbots.

Companies are increasingly looking at AI agents that can access business information, understand context and perform actions.

Salesforce's AIForce represents one approach to this transition: instead of replacing its existing platform, Salesforce wants its data, workflows and business logic to become accessible through a growing range of AI interfaces.

This could eventually change how employees use CRM and business software.

However, the success of this approach will depend on several factors, including security, reliability, data governance, integration with existing systems and whether businesses trust AI agents to perform important tasks.

Conclusion

Dreamforce 2026 has become more than a product launch event for Salesforce.

Inside the Moscone Center, Marc Benioff presented AIForce as a new way to connect Salesforce’s data, workflows and AI agents with different interfaces. The company also highlighted its broader Agentforce strategy and partnerships involving technologies such as Anthropic’s Claude.

Outside the venue, protesters raised questions about Salesforce’s government contracts and demanded an end to relationships involving ICE.

Meanwhile, the appearance of AI leaders including Dario Amodei and Jensen Huang highlighted the growing disagreement over AI safety, regulation and the pace of technological development.

Together, these developments show the two sides of the current AI transformation: the race to make AI more capable and useful, and the growing debate over how that technology should be governed and deployed.

For enterprise technology, Dreamforce 2026 therefore represents not just another Salesforce product cycle, but a broader shift toward an AI-driven software ecosystem in which humans, business applications and autonomous AI agents increasingly work together.