Salesforce Inc. is cranking out yet more artificial intelligence agents, this time for its flagship customer relationship management platform.
The new bots include the Einstein Sales Development Rep, or Einstein SDR, and the Einstein Sales Coach Agent, and they’re designed to aid sales representatives in a variety of ways.
The Einstein SDR Agent is designed to handle sales prospects and sort the wheat from the chaff. What’s unique about this agent is that it possesses “agentic AI” capabilities that enable it to take actions on behalf of their users. For instance, it has the ability to engage autonomously with new sales prospects and then make decisions or prioritize actions that align with the organization’s goals.
It does this by using retrieval-augmented generation techniques to analyze a sales prospect’s question, so it leverages the company’s knowledge base to determine what should be done next – such as answering questions about a product, handling objections or scheduling a meeting with a sales rep. The idea is to help filter out the prospects and send the most promising ones to human agents, so they can close the deal.
Salesforce added that the Einstein SDR Agent has been designed to minimize the need for human agents to perform actions when new leads appear. For instance, if a new prospect visits the company’s website and downloads a white paper, Einstein SDR Agent will create an email and reach out to that lead by itself. It will then handle the communications in case that lead responds, with the goal being to set up a meeting with the human agent.
As for the Einstein Sales Coach Agent, this is all about coaching sales representatives through role-playing scenarios. It works by facilitating lifelike situations that are tailored for each sales lead, giving agents the opportunity to practice and increase their chances of obtaining a successful outcome.
In each role-play, Einstein acts as the sales prospect, and once again it relies on RAG techniques to try to replicate how that prospective buyer is likely to respond, based on the direction the conversation goes. Each response is grounded in data stored within the Salesforce CRM, such as information regarding the deal in question, the customer’s account, previous correspondence and so on.
Salespeople can use the Einstein Sales Coach Agent to practice their pitches and the way they carry out negotiations, prior to doing so in real life, and perhaps improve their chances of achieving a successful outcome.
Once the role-playing scenario concludes, Einstein Sales Coach Agent will provide personalized and objective feedback to the salesperson, based on their performance. That includes hints and tips on what they could have done to improve the chances of success.
In addition, sales team managers will be able to connect the coaching sessions to the salesperson’s actual results, and in that way they can assess its overall impact.
Salesforce explained that the new models are built on the Einstein 1 Agentforce platform and can quickly be set up using no-code actions. They’ll become generally available to Salesforce CRM customers in October.
Once they’re live, users will be able to create workflows using prebuilt templates, based on Salesforce’s proven best practices for selling. To further enhance the models, customers can upload relevant external information to the Salesforce Data Cloud, such as their existing selling and training documentation. This will provide the agents with more accurate information to improve their outputs and make them more realistic, the company said.