Salesforce has unveiled a ‘first-of-its-kind’ Agentforce Testing Center for AI agents.
Announced last week, the offering boasts New Data Cloud Sandboxes that enable teams to test their AI and data lifecycles using synthetically generated data.
Moreover, users may check their agent responses and actions are correct while providing complete monitoring of usage and feedback at scale.
Adam Evans, EVP and GM for the Salesforce AI Platform, explains how and why the Agentforce Testing Center is such a useful toolkit for enterprises:
Agentforce is helping businesses create a limitless workforce. To deliver this value fast, CIOs need new tools for testing and monitoring agentic systems.
“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 Center, which will help companies roll out trusted AI agents with no-code tools for testing, deploying, and monitoring in a secure, repeatable way.”
The CRM leader also suggests that other vendors are unable to offer their customers the necessary testing environment for their AI creations before deploying them.
As a result, the AI may cause hallucinations, inaccurate results, and poor customer experiences.
New Capabilities
Salesforce outlines four central features of the AI testing solution, which include the automated testing function, sandboxes for Agentforce and Data Cloud, Agentforce monitoring, and usage insights in the Digital Wallet.
Using natural language instructions, the Testing Center can automatically generate hundreds of test interactions like customer requests with the Agentforce Service Agent. The results can be tested in parallel to see how often it achieves the correct response.
This data can then be used to improve the instructions to make the expected topic selected more often, thereby enhancing the customer experiences that follow.
Salesforce Sandboxes can replicate your organization’s data and metadata into a safe environment in which development teams can quickly bring together unstructured data foundations and testing in Agentforce without risking damaging their business reputation.
Teams can also perform User Acceptance Testing (UAT) with a group of users as a further precaution and then integrate those changes into production with DevOps Center, Change Sets, and the Salesforce CLI that supports Data Cloud and Agentforce.
The Einstein Trust Layer can be tested in pre-production following the release of Data Cloud Sandboxes. Einstein’s audit trail and feedback store allow teams to build closed loops for AI testing.
The new observability solutions, Agentforce and Utterance Analysis offer insights into the adoption and accuracy of capabilities while Agentforce is live.
Finally, Data Cloud Sandbox and Agentforce usage are metered in the Digital Wallet so that customers can have visibility into consumption through the AI development lifecycle. Granular insights have been made possible by the addition of new improvements so that teams can isolate new trends relating to usage.
Due to the Wallet’s integration into the Salesforce platform, teams can also create automated alerts to let admins know if usage surpasses a particular level
Keith Kirkpatrick, Research Director at The Futurum Group, shared his analysis of the CRM vendor’s latest updates: “To help engender trust in AI, every business investing in an agent strategy will need to think through how they pressure test an autonomous AI agent that can reason, retrieve data, and use tools.
Salesforce is instilling confidence by using AI to test hundreds of variations of an agent’s interaction in parallel.
“With these enhancements and support for Agentforce and Data Cloud in familiar testing environments, this Testing Center will help Salesforce customers manage the unique needs of testing AI agents.”
Earlier this month, Salesforce completed the acquisition of Own Company to strengthen its Data Cloud’s security and protection offering.
Having initially been announced in September, the $1.9BN deal is now complete.
The roll-up marks Salesforce’s most expensive purchase since Slack for $28BN in 2021.
