At Dreamforce 2026, Salesforce unveiled AIforce alongside Koa and Agentforce Coworker, extending enterprise data, reasoning, and workflows across AI-driven interfaces.
Dataquest Bureau19 Sep 2026
09:01IST
New Update
Salesforce arrived at Dreamforce 2026 with a message that went beyond a product launch. It declared that the era of employees opening an app, searching records, and clicking through dashboards was ending. In its place, the company unveiled AIforce, a new interface layer built to carry Salesforce data, workflows, business logic, and governance directly into the AI platforms people already use every day.
“AI is creating an interface revolution,” said Marc Benioff, Chair and CEO, Salesforce. “We are combining model intelligence with all the context that customers have built into Salesforce to create an intelligent, dynamic, composable system that is securely governed, built with Zero Data Retention, and designed to work with the core systems that already run your business.”
A more personalised way to work
A large amount of a company’s knowledge already existed inside Salesforce, including customer information, conversations, workflows, business rules, and operational history. Until then, employees generally needed to enter Salesforce to access this information.
AIforce changes that experience by allowing users to ask questions, retrieve relevant information, update records, and trigger approved workflows through an AI interface. It also introduced dynamic and personalised interfaces, allowing users to describe what they needed instead of working only with fixed dashboards, layouts, and fields.
These experiences could be created within platforms such as Slack, Claude, and Agentforce Coworker, allowing enterprise software to adapt more closely to the way people worked.
The backbone of an agentic enterprise
AIforce becomes an important part of Salesforce’s wider Agentic Enterprise strategy. At its foundation was Data 360, which brought together harmonised and federated data, metadata, and memory so AI agents could understand customer and business context.
Customer 360 provided the business logic, processes, permissions, and actions required across sales, service, marketing, commerce, and other functions. Agentforce provided Salesforce’s digital workforce, including ready-to-use agents and tools for companies to build their own.
AIforce connected these capabilities to the different AI interfaces where employees and agents worked, creating a more connected way to access and use enterprise intelligence.
Salesforce met Claude
Through Claudeforce, Salesforce extended its partnership with Anthropic by bringing Salesforce intelligence directly into Claude. Salesforce in Claude used a prebuilt Model Context Protocol (MCP) server, reducing the complexity normally involved in connecting enterprise data and AI models.
The experience had been piloted by organisations including Deloitte, GitLab, and Legora and was made available to customers in beta. Salesforce also announced plans to extend the offering with Tableau analytics and capabilities covering service, marketing, commerce, and industry-specific use cases.
CRM moved into the conversation
With Slackforce, Salesforce brought its customer relationship management (CRM) intelligence into Slack, where employees already communicated and collaborated.
Slackforce surfaces enabled teams to bring live information from Salesforce, Slack, and other tools into interactive interfaces that they could explore, filter, discuss, and act on together.
Salesforce also introduced Slack CRM, allowing users to perform common CRM activities through simple prompts within Slack, including creating accounts, recording notes from customer calls, and updating Salesforce records. This reduced the need to constantly switch between collaboration and business applications.
Agentforce Coworker as an AI teammate
Salesforce also expanded Agentforce through Agentforce Coworker, an AI teammate that worked directly inside Salesforce Lightning. It could analyse accounts, activities, and historical information, surface relevant insights, and take approved actions while operating within existing Salesforce permissions and business rules.
Salesforce said 100,000 users had activated Coworker within its first 35 days.
Governance stayed front and centre
As AI moved from simply answering questions to taking actions, security and governance became increasingly important for enterprises.
Salesforce said AIforce operated within existing Salesforce permissions and business rules, meaning agents were expected to access only information that users were authorised to see, while actions were routed through Salesforce.
The company also highlighted zero data retention, stating that business data used to answer a request was not retained by the model provider.
A new customer service agent
Salesforce also introduced Fin, an industry-leading customer service agent platform from the company it recently acquired. Fin strengthened Salesforce’s Agentforce service capabilities by expanding its AI-powered customer support offerings and adding a global customer base of more than 30,000 companies into the Salesforce ecosystem.
Salesforce and NVIDIA introduced Koa
Salesforce and NVIDIA also announced Koa, Salesforce’s first CRM reasoning model for Agentforce, built on NVIDIA Nemotron.
Koa was designed to help AI agents manage complex, multistep workflows and select the right tools to complete tasks. Salesforce said the model was developed by post-training NVIDIA Nemotron 3 Super using a proprietary synthetic dataset based on enterprise knowledge from nearly three decades of CRM deployments.
The objective was to create a reasoning model that understood not only language but also how businesses operated, including their workflows, processes, and operational policies.
“AI is creating a much larger opportunity for software. Every company needs useful AI, tailored to its knowledge, expertise, and work. NVIDIA Nemotron open models give Salesforce the foundation to turn decades of enterprise expertise into specialised AI with Koa, creating a CRM model that can reason and securely take action,” said Jensen Huang, founder and CEO of NVIDIA.
The bigger shift
Taken together, the announcements at Dreamforce 2026 pointed to something larger than a single product cycle. Salesforce brought together Data 360 for context, Customer 360 for process and intelligence, Agentforce for autonomous action, Koa for specialised reasoning, and AIforce to deliver all of it through the interfaces people already trusted.
The underlying bet was simple: employees should no longer need to go to their business software to get work done. Instead, enterprise intelligence would come to them, inside the tools they already opened every day.
If that vision held up at scale, the next era of enterprise software would be defined less by which application someone opened and more by how seamlessly AI could understand a business, respect its rules, and help its people move faster.
By Shipra Sinha, Senior Analyst, CMR
