Salesforce has taken its latest step into the AI space with Agentforce, a platform for implementing generative AI in areas such as customer service, service, sales and marketing. It will be available on October 25.
In a press conference on September 12, Salesforce CEO Marc Benioff emphasized that Agentforce customers won't need to build their own AI models or even manually choose between models and that AI won't be an “extra co-pilot” that adds overhead to existing services.
Services formerly called Einstein Copilot, including Einstein Copilot in Slack, will now be branded under the Agentforce brand.
Agentforce turns chatbots into generative conversationalists
Agentforce is a deployment platform for generative AI agents that can autonomously interact with current and potential customers. It connects to the Salesforce data cloud and filters structured and unstructured data through the Atlas Reasoning Engine, a pillar of the company’s generative AI deployment.
Data Cloud’s data lake can store “zero-copy” data, pulling it from external systems without duplicating it. The Atlas reasoning engine is a proprietary tool that Salesforce has not detailed in detail. However, they did explain that Atlas serves as a mechanism to:
- Filtering AI agents' decisions.
- “Reasoning” which data is most relevant.
- Contrast the proposed answers with real-world information.
Agentforce agents can hold autonomous conversations with customers to answer service questions, call potential customers, or work on marketing campaigns. They can interact with customers via voice, phone, or text. During the September 12 event, Salesforce demonstrated how an AI “personality” — with its own name, access to company information, and settings — can be used multimodally across multiple media.
SEE: What are the different stages of CRM?
Benioff recommended using Agentforce as a replacement for Tier 1 customer service staff, the front-line responders. AI agents could still escalate more difficult questions to Tier 2 workers, the more experienced staff members.
Agentforce can be implemented as a low-code solution, where a library of agent actions can be integrated into a menu, with some agent customization done using natural language.
Prompt Builder, a service for creating Data Cloud prompts, helps users customize autonomous agents. A low-code template builder is available for organizations that want to use their preferred APIs and LLMs in Prompt Builder.
Agents will be subject to some of the same security and call monitoring standards as human agents. The existing Salesforce Omni Supervisor feature will be able to monitor agents and flag issues. Salesforce AI CEO Clara Shih said this can allow human supervisors to intervene in case information coming from Agentforce is inappropriate or inaccurate.
“Omni Supervisor is now a place where service managers can also manage their AgentForce agents. They can filter by topic, they can spot trouble spots…” Shih said. “Just like with a new employee with a human, they can step in and make sure nothing goes wrong.”
Salesforce tries to solve the problem of phone trees and wait times
According to Benioff, the current era of AI has left organizations wondering where tangible customer value lies.
Customers “want their time back,” Benioff said, positioning Agentforce as a potential alternative to waiting in line with a customer service representative.
Agentforce will be priced per conversation
Agentforce costs $2 per conversation. Salesforce did not detail how the $2 per conversation cost compares to human workers.
Agentforce for Service and Sales will be available on October 25, but certain elements of the Atlas Reasoning Engine will not be ready until February 2025.
Salesforce Product Releases in October
Additional Salesforce updates coming in the October update include a quota attainment planning tool and ServiceCloud Employee Services, an AI-powered solution for employee support.
