Salesforce Focuses Dreamforce on Outcome-Based Pricing and Agent Governance

Enterprise artificial intelligence deployments are shifting focus from conversational demonstrations to measurable operational outcomes, according to reports from Salesforce Inc.’s Dreamforce 2026 conference in San Francisco by SiliconANGLE . Rather than isolated toolkits built for rapid proofs of concept, enterprise technology teams are seeking integrated architectures that link automated agents directly to corporate data, existing workflows, and human managers.

The demand for measurable return on investment, task-level governance, and structured digital labor management shaped Salesforce's latest product rollouts and its AIforce architectural layer. Speaking with theCUBE during the event, George Gilbert, principal analyst for data platforms, intelligent apps, and agentic frameworks at theCUBE Research, noted that Salesforce has integrated the underlying components so organizations can move past basic request-and-response interactions to automated outcomes across data, applications, and agents.

To align billing with functional performance, Salesforce now charges for Agentforce service agents only when a customer issue is resolved rather than when a task escalates to a human agent. Built-in analytics allow organizations to adjust agent behavior as business workflows evolve, according to Kishan Chetan, executive vice president and general manager of Agentforce and Service Cloud at Salesforce.

Early adopters described assigning autonomous agents to structured corporate roles with direct management oversight. Brandon Metcalf, chief executive officer of software firm Asymbl Inc., reported that his company uses digital workers—including an automated recruiting agent that assisted in hiring approximately 100 people across 100 days—under human managers with explicit performance metrics. Metcalf stated the productivity impact of those digital workers rose from roughly $5 million last year to $13 million this year.

Within Salesforce's collaboration portfolio, Slackbot has been adapted to function as a Model Context Protocol client. The tool connects to authorized enterprise agents, retrieves company data, and generates shareable interfaces on demand. Ryan Gavin, executive vice president and chief marketing officer of Slack at Salesforce, noted that the tool has become the fastest-adopted product in company history, reaching 1.5 million monthly active users since January.

Salesforce also detailed its Enterprise AI Harness, an open composable platform combining Data 360, MuleSoft, and Agentforce. The platform restricts agents to task-level policies defined by administrators. Savinay Berry, executive vice president and general manager of data management at Salesforce, emphasized that granular policy controls are critical as enterprise systems approach generating an estimated 150 to 200 zettabytes of data.