Understand this story
On October 27, 2026, Salesforce announced the addition of Agent Evals, API Evals, and advanced cost controls to Agent Fabric, its AI control plane. According to Unite.AI, Agent Evals automatically generates tests before deployment and evaluates agents, LLMs, and MCP servers in production against quality thresholds, triggering preset interventions including the Agent Kill Switch when standards are not met; the FinOps Dashboard and Model Wallet provide spending caps by agent, instance, or business unit, and Cost Optimization Policies can reduce token usage by up to 75% depending on the test scenario. The report states that the platform handles more than 1.2 billion requests per day, and that Agent Evals and API Evals entered beta in November 2026 and are planned for general availability in January 2027.
AI-generated from public reports
Latest development · Report
Salesforce Adds Agent Evals and Cost Controls to Agent FabricDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 27
Unite.AI
Salesforce Adds Agent Evals and Cost Controls to Agent Fabric
On October 27, 2026, Salesforce announced the addition of Agent Evals, API Evals, and advanced cost controls to its AI control plane Agent Fabric; the platform handles more than 1.2 billion requests per day. Agent Evals automatically generates tests before deployment and evaluates agents, LLMs, and MCP servers in production against quality thresholds, triggering preset interventions including the Agent Kill Switch when standards are not met. The FinOps Dashboard and Model Wallet provide spending caps by agent, instance, or business unit, and Cost Optimization Policies can reduce token usage by up to 75% depending on the test scenario. Agent Evals and API Evals enter beta in November 2026 and are planned for general availability in January 2027.