Salesforce announced new agent observability tools, bringing deeper visibility, monitoring, and optimization to every AI agent within the Agentforce 360 Platform.
By providing granular insight into agent behavior, the tools empower organizations to safely scale AI and continually improve agent performance. And by fostering better collaboration between humans and intelligent agents, these observability tools ensure that AI adoption is both reliable and trustworthy, accelerating every company’s journey to becoming an Agentic Enterprise.
The new observability tools within Agentforce span three core areas: analytics, optimization, and health monitoring.
Agent Analytics provides a comprehensive view of how every agent is performing, translating performance into meaningful data, trends, and insights.
- Track Agent Performance: Monitor usage and effectiveness metrics across all deployed agents, giving teams a clear picture of how agents perform in real customer interactions.
- KPI Trends: Surface KPI trends over time, making it easy to see where performance is improving, declining, or requiring attention.
- Actionable Insights: Address ineffective topics, actions, or flows, enabling teams to take targeted steps to optimize agent performance.
Agent Optimization gives customers full observability into every agent interaction. By illuminating performance gaps, tracing session flows, and revealing exactly how agents make decisions, businesses can quickly diagnose issues, understand why an agent behaved a certain way, and take targeted actions to continuously correct and improve outcomes.
- Observe Every Interaction: Access end-to-end visibility of every Agentforce interaction to see exactly how agents respond, step by step, even across complex reasoning chains.
- Cluster and Analyze Sessions: Group similar requests to uncover patterns, friction points, and quality trends while scoring agent responses using intent, topic, and quality metrics.
- Optimize Agent Configuration: Identify configuration issues affecting performance and understand exactly where tuning, retraining, or guardrails are needed.
Agent Health Monitoring ensures continuous agent uptime, reliability, and responsiveness by providing near-real-time visibility and actionable trust signals on your deployed fleet. This component is essential for operational rigor, ensuring your agents perform reliably and consistently even during peak loads.
- Monitor Agent Status Continuously: Track key health metrics in near real time, ensuring dashboard data is fresh and surfacing potential issues before they impact performance or trust signals.
- Resolve Failures Proactively: Receive high-speed alerts on critical errors, latency spikes, and escalations, enabling teams to quickly detect, investigate, and resolve problems to minimize downtime.
- Maintain Enterprise Reliability: Leverage a system built for high availability and operational robustness, giving you confidence that continuous monitoring scales reliably with your entire agent fleet.
Agentforce 360 is powered by two foundational building blocks:
- Session Tracing Data Model: This data model logs every interaction, including user inputs, agent responses, reasoning steps, LLM calls, and guardrail checks, and stores them securely in Data 360. This foundation provides unified visibility and granular, session-level insights so you can ensure agents are behaving as intended and responding appropriately.
- MuleSoft Agent Fabric: This new capability provides a single place to register, manage, govern, and observe every agent, regardless of where it was built.
With these new agent observability tools, plus robust governance in MuleSoft and unified context in Data 360, every company can get full end-to-end control of all their agents, all from within the Agentforce 360 Platform.
Deep observability in Agentforce Studio, including Agent Analytics and Agent Optimization, is available now.
Agent Health Monitoring will be generally available in Spring 2026.
Regional rollout for Agent Optimization and Agent Analytics: EMEA customers is on November 20, 2025, and APAC customers on November 21, 2025.
The Latest
80% of respondents agree that the IT role is shifting from operators to orchestrators, according to the 2026 IT Trends Report: The Human Side of Autonomous IT from SolarWinds ...
40% of organizations deploying AI will implement dedicated AI observability tools by 2028 to monitor model performance, bias and outputs, according to Gartner ...
Until AI-powered engineering tools have live visibility of how code behaves at runtime, they cannot be trusted to autonomously ensure reliable systems, according to the State of AI-Powered Engineering Report 2026 report from Lightrun. The report reveals that a major volume of manual work is required when AI-generated code is deployed: 43% of AI-generated code requires manual debugging in production, even after passing QA or staging tests. Furthermore, an average of three manual redeploy cycles are required to verify a single AI-suggested code fix in production ...
Many organizations describe AI as strategic, but they do not manage it strategically. When AI plans are disconnected from strategy, detached from organizational learning, and protected from serious assumptions testing, the problem is no longer technical immaturity; it is a failure of management discipline ... Executives too often tell organizations to "use AI" before they define what AI is supposed to change. The problem deepens in organizations where strategy isn't well articulated in the first place ...
Across the enterprise technology landscape, a quiet crisis is playing out. Organizations have run hundreds, sometimes thousands, of generative AI pilots. Leadership has celebrated the proof of concept (POCs) ... Industry experience points to a sobering reality: only 5-10% of AI POCs that progress to the pilot stage successfully reach scaled production. The remaining 90% fail because the enterprise environment around them was never ready to absorb them, not the AI models ...
Today's modern systems are not what they once were. Organizations now rely on distributed systems, event-driven workflows, hybrid and multi-cloud environments and continuous delivery pipelines. While each adds flexibility, it also introduces new, often invisible failures. Development speed is no longer the primary bottleneck of innovation. Reliability is ...
Seeing is believing, or in this case, seeing is understanding, according to New Relic's 2025 Observability Forecast for Retail and eCommerce report. Retailers who want to provide exceptional customer experiences while improving IT operations efficiency are leaning on observability ... Here are five key takeaways from the report ...
Technology leaders across the federal landscape are facing, and will continue to face, an uphill battle when it comes to fortifying their digital environments against hostile and persistent threat actors. On one hand, they are being asked to push digital transformation ... On the other hand, they are facing the fiscal uncertainty of continuing resolutions (CR) and government shutdowns looming near and far. In the face of these challenges, CIOs, CTOs, and CISOs must figure out how to modernize legacy systems and infrastructure while doing more with less and still defending against external and internal threats ...
Reliability is no longer proven by uptime alone, according to the The SRE Report 2026 from LogicMonitor. In the AI era, it is experienced through speed, consistency, and user trust, and increasingly judged by business impact. As digital services grow more complex and AI systems move into production, traditional monitoring approaches are struggling to keep pace, increasing the need for AI-first observability that spans applications, infrastructure, and the Internet ...
If AI is the engine of a modern organization, then data engineering is the road system beneath it. You can build the most powerful engine in the world, but without paved roads, traffic signals, and bridges that can support its weight, it will stall. In many enterprises, the engine is ready. The roads are not ...