Salesforce launches AIforce at Dreamforce '26: the CRM leader bets big on the agentic enterprise

Salesforce has never been shy at Dreamforce, and its 2026 event in San Francisco (15–17 September) delivered its boldest AI statement yet. The headline was AIforce, a new AI interface layer that sits above Salesforce's core products and lets people get work done by talking to AI rather than clicking through screens. For the world's leading CRM, it's a confident, clarifying move, and a strong signal of where enterprise software is heading.

What Salesforce announced

AIforce is best understood as a "live interface layer" over the Salesforce stack, uniting Agentforce, Data 360 and Customer 360 so AI agents can reason across your data and take real actions inside your existing security model. You can update records, trigger workflows and answer questions without opening a traditional Salesforce dashboard, work simply happens where you already are.

CEO Marc Benioff framed it as an "interface revolution", combining model intelligence with the context and business logic that customers have spent years building inside Salesforce. As Salesforce's Patrick Stokes put it, the value of the platform "has never been in the UI", it's in the trusted data and logic underneath. AIforce leans into exactly that.

The launch came with several concrete pieces:

• Claudeforce brings Salesforce directly into Claude, shipping with 37 prebuilt sales skills (now in open beta).
• Slackforce surfaces Salesforce context inside Slack, so you can create records and act on data straight from a conversation.
• Agentforce Coworker, an AI teammate that works alongside you in the Lightning experience, available now.
• A headless toolkit that exposes Salesforce capabilities through MCPs, APIs and plugins, so agents and apps can plug into the platform anywhere.

Crucially, all of it runs within existing permissions and enforces zero data retention, a governance-first posture that matters enormously to enterprise and regulated customers.

Momentum behind the message

This wasn't a vision-only keynote. Salesforce says 30,000 customers are now live on Agentforce, real production adoption, not pilots, which is a remarkable number for a platform still early in its life. The company also introduced Koa, its first CRM reasoning model, built on NVIDIA Nemotron, and rolled out a wave of specialist agents (Casey, Paige, Carter, Marshall, Piper and Fin are generally available, with Hunter in pilot ahead of a November 2026 release).

Taken together, it's a coherent story: trusted data, a reasoning model tuned for CRM, a fleet of ready-made agents, and now an interface layer that makes all of it accessible wherever people work.

Why this cements Salesforce's lead

Plenty of vendors are shipping AI features. What sets this apart is that Salesforce is reorganising its entire product philosophy around agents while keeping its biggest advantages intact, the depth of customer data, the governance model, and the business logic that large organisations already trust. "AI replaces the UI" is a genuinely ambitious idea, and Salesforce is one of the few companies with the data gravity and install base to actually pull it off.

For a CRM already regarded as the market leader, AIforce reframes the category on Salesforce's terms: the winner won't be whoever bolts on the most chatbots, but whoever can let AI act safely across trusted enterprise data. That's a strong position to be arguing from.

What it means for your business

The exciting part is also the part that needs a clear head. An interface layer that lets agents act across your CRM is powerful precisely because it can do things, create records, trigger workflows, move data, so the value and the risk both scale with how well it's set up. The organisations that win with AIforce will be the ones who get the foundations right:

• Clean, well-governed data, because agents are only as good as the Customer 360 and Data 360 underneath them.
• Permissions and guardrails, so an agent can act quickly without acting outside its lane.
• A clear adoption plan, starting with a high-value workflow (sales follow-ups, service triage, quoting) and measuring the result before expanding.

This is where an independent advisor earns its keep. As a vendor-neutral partner, we help you separate genuine capability from launch-day excitement, design the governance so AI is an asset on your risk register rather than a liability, and make sure a platform as capable as Salesforce is configured to deliver real, measurable value for your business.

Salesforce has made its direction unmistakable, and it's an impressive one. If AI-powered CRM is on your roadmap, now is the right time to get your data, governance and adoption strategy ready to make the most of it.

Sources: Salesforce Ben, "Salesforce Launches AIforce at Dreamforce '26"; Salesforce Ben, Agentforce keynote roundup; Salesforce, Dreamforce.