2 min read

Salesforce Doubles Down on AI with Agentforce 2.0

Just two months after launching Agentforce, its generative AI tool for enterprises, Salesforce is set to unveil Agentforce 2.0, a significantly enhanced version, reports SiliconANGLE. This upgrade promises to revolutionize digital labor, building on Salesforce’s vision of AI-powered agents that augment human employees.

Agentforce enables companies to construct and customize AI agents, which operate autonomously to support various tasks. These agents, often referred to as "workers," leverage pre-built automations and business logic to handle tasks across sales, service, marketing, and commerce.

Salesforce CEO Marc Benioff has termed this approach the "agent revolution," aiming to redefine how digital labor functions within organizations.

Agentforce 2.0 introduces the ability to combine multiple skill templates, creating agents that can perform a wider range of tasks. For example, a service agent could now handle both customer service calls and generate marketing profiles, streamlining customer interaction and producing valuable marketing analytics.

"We’ve got thousands of partners building their skills," said Adam Evans, Salesforce's executive vice president of Salesforce AI, to SiliconANGLE. "There’s a really wide gamut of the jobs that they can do when you start mixing these different skills."

Salesforce has expanded the library of pre-built agent skills, making them accessible through the Salesforce AppExchange and encompassing CRM, Slack, Tableau, and external partner integrations. The integration with Slack is particularly noteworthy, allowing agents to interact directly within the communication platform, streamlining workflows.

Agentforce 2.0 also provides users with greater control over agent customization and testing.

"We’re building the tools to allow people to bring multiple skills together quickly," Evans explained. "We’re also building the tools to launch things into a testing center and more. So that you can say, here are the use cases I care about, test it."

The Agent Builder, Salesforce's AI agent customization engine, now utilizes Agentforce to guide users in creating agents. This tool allows users to describe their desired agent characteristics and objectives in natural language, receiving recommendations for custom skill sets. This feature also facilitates the design of cross-departmental agents, simplifying workflows.

Evans acknowledged that AI agents are still in their early stages of development, emphasizing the importance of careful testing to ensure agents are not overloaded with unnecessary skills. Salesforce has introduced an AI playground and testing center to enable thorough evaluation before deployment.

Agentforce 2.0 is built upon Salesforce's Atlas reasoning engine, which aims to replicate human-like thinking. Salesforce has invested in enhancing Atlas's capabilities, resulting in significant accuracy improvements in real-world scenarios.

"Results from our customers in field testing show it’s about twice as accurate than other techniques," Evans stated, comparing Atlas to vector databases and other solutions.