November 20, 2024 –Salesforce (NYSE: CRM), the world’s #1 AI CRM, today announced agentic lifecycle management tools to automate Agentforce testing, prototype agents in secure Sandbox environments, and transparently manage usage at scale.
AI agents are a new paradigm in software. They are intelligent systems that can reason and act on behalf of customers and employees. But to realise their full potential, agents need to be tested and configured without disrupting live production environments. This new toolchain — the first of its kind in the industry — will enable teams to test, deploy, and monitor AI agents with Agentforce at scale, with confidence, enabling every enterprise to become “agent-first.”
“Agentforce is helping businesses create a limitless workforce. To deliver this value fast, CIOs need new tools for testing and monitoring agentic systems,” said Linda Saunders, Salesforce Director Solutions Engineering Africa. “Salesforce introduced the concept of Application Lifecycle Management back in 2006 with Force.com. This new category of Agentic Lifecycle Management requires unique tools, and Salesforce is meeting the moment again with Agentforce Testing Centre, which will help companies roll out trusted AI agents with no-code tools for testing, deploying, and monitoring in a secure, repeatable way.”
Other vendors lack the necessary capabilities for customers to run the appropriate tests on their AI before deploying, which can lead to hallucinations, inaccurate results, and substandard customer experiences. Agentforce Testing Center — built on the enterprise-grade Salesforce Platform and integrated with Data Cloud — enables every organisation to easily test and monitor AI agents so that they can deploy with confidence.
New capabilities include:
● AI-generated tests for Agentforce: Teams building with Agentforce need to accurately test all of the different ways a customer may pose a question or interact with an agent. In addition to Agent Builder, which features a Plan Tracer for investigating the reasoning process of an agent, the new Agentforce Testing Center enables teams to test topic and action selection at scale. Using natural language instructions, Testing Center can auto-generate hundreds of synthetic interactions — such as requests a customer may make when engaging with Agentforce Service Agent — then test them in parallel to see how frequently they result in the right outcome. Teams can then use the test data to refine instructions so the expected topic is more frequently selected, improving the end customer experience.
● Sandboxes for Agentforce and Data Cloud: Teams looking to test Agentforce need to do so in safe, isolated environments. Generally available today, Salesforce Sandboxes — mirror images of your production org’s data and configurations — now support both Data Cloud and Agentforce. By replicating the org’s data and metadata into a risk-free environment, development teams can rapidly assemble their unstructured data foundation and rigorously prototype Agentforce without fear of disrupting the business. Now teams can perform UAT (User Acceptance Testing) with an initial set of users to ensure that Agentforce performs the tasks that it’s intended to accomplish, then migrate those changes to production using familiar tools such as Change Sets, DevOps Center, and the Salesforce CLI that now support Data Cloud and Agentforce.
● Monitoring and observability for Agentforce: With the general availability of Data Cloud Sandboxes, the full Einstein Trust Layer can be tested in a secure, pre-production environment, enabling rapid configuration of Agentforce agents and Prompt Templates. With the Einstein Trust Layer’s audit trail and feedback store in sandboxes, teams can build a closed loop for AI testing — iterating on prompts and actions based on user feedback. And once Agentforce is live in production, new capabilities for granular insights into adoption and accuracy become available through Agentforce Analytics and Utterance Analysis – new observability solutions built natively on Data Cloud for continuous iteration while moving through the Agentforce lifecycle.
● Transparent usage monitoring in Digital Wallet: Data Cloud Sandbox and Agentforce usage is metered in Digital Wallet, providing customers with complete visibility into their consumption across the AI development lifecycle. New enhancements provide granular insights into what features consume credits, so that teams can uncover new trends around usage as they scale.
And because Digital Wallet is integrated into the Salesforce Platform, teams can create automations to alert admins if, for example, usage exceeds a particular threshold. “From the documents we have, it is clear the company directors are the two sons of the DP and that is the nexus between the DP and that particular company,” Mohamud told the Senate.
