Salesforce is entering the AI agent landscape with the launch of its innovative Einstein Service Agent, which leverages the company’s established Einstein AI platform.
The Einstein Service Agent offers a generative AI-driven self-service experience designed for end customers. This conversational AI interface enables users to resolve various issues and answer questions effectively. Built on a foundation similar to Einstein Copilot—an internal tool for employees—this new service enhances customer interactions.
The core concept of agentic AI is to move beyond simple Q&A to perform actions for users. For instance, the Einstein Service Agent can facilitate product returns or process refunds. It is capable of managing a range of interactions, from simple inquiries to complex multi-step workflows, all integrated within Salesforce’s approved company processes.
Initially, the Einstein Service Agent will focus on customer service applications, with plans to expand to other Salesforce clouds in the future.
What sets Einstein Service Agent apart from competitors is its seamless integration with Salesforce’s existing customer data and workflows. Clara Shih, CEO of Salesforce AI, emphasized, “Einstein Service Agent is a generative AI-powered, self-service conversational experience grounded in our trust layer and platform. It utilizes customer data and official workflows developed over the last 25 years.”
Understanding the Difference: AI Agent vs. AI Copilot
Over the past year, Salesforce has been unveiling various components of its generative AI strategy. Notably, the Einstein Copilot, which became generally available in April, facilitates a wide variety of conversational AI experiences for direct Salesforce users.
“Einstein Copilot is employee-facing, tailored for sales teams, customer service representatives, marketers, and knowledge workers,” Shih explained. “In contrast, Einstein Service Agent is designed for our customers’ customers, enhancing their self-service capabilities.”
While conversational bots have been around for a while, Shih clarified that Einstein Service Agent distinguishes itself by utilizing Salesforce’s extensive data and AI advancements. It’s not just about answering basic questions; it enables customers to receive knowledge-driven responses and take action.
The copilot can integrate multiple AI engines and responses, similar to the AI agent approach. Shih characterizes the differences as a matter of semantics and autonomy.
“This is a spectrum toward greater autonomy,” Shih noted.
Driven by Existing Customer Workflows
As an example of practical application, Shih mentioned that Salesforce is collaborating with a major apparel company to pilot the Einstein Service Agent.
When a customer places an online order and receives the wrong product, they can contact a human agent during business hours. That agent may use the Einstein Copilot for assistance. However, for after-hours issues or customers seeking a completely self-service experience, the Einstein Service Agent will facilitate resolution.
With the Einstein Service Agent, customers can inquire about their issues and, depending on the established workflow, receive immediate resolutions.
The workflow leverages existing data within Salesforce's Service Cloud, allowing the Einstein Service Agent to tap into pre-defined protocols for effective problem resolution. Shih highlighted that Einstein Studio manages the administrative and configuration aspects of all Einstein AI initiatives, including Service Agents, enhancing customer engagement through established logic.
“All that our company has invested in over the past 25 years is now empowering us to offer customers secure and trusted access to generative AI,” Shih concluded.