SAN FRANCISCO, April 29, 2024 — Selling power announced the general availability of Einstein Copilot, the AI conversational assistant for CRM, along with new features designed to improve salesperson productivity and expand the adoption of generative AI.
For 25 years, Salesforce has helped sales teams find, win, and retain customers across all facets of the sales cycle – from prospecting and activity tracking to forecasting and revenue operations – all of which can be tailored to suit your needs. customer activity through customizable metadata. Salesforce uses this wealth of experience to transform the most common and valuable sales processes into conversational AI interactions powered by an organization’s business-specific data and metadata, and delivered at scale through new analysis and testing tools.
A key part of this metadata, and unique to Einstein Copilot, are Copilot Actions – pre-programmed features that allow Einstein Copilot not only to answer questions using business data, but also to chain workflows together to get things done on behalf of users. In fact, in the recent “Your Data, Your AI” survey, more than half of workers said that AI based on private company data increases their trust in the tool.
“As we continue the rapid pace of innovation, this is an exciting advancement in conversational AI for businesses,” said Adam Evans, senior vice president of products at Salesforce AI Platform. “With this release, every organization can now deploy a trusted AI assistant, powered by their unique business data and metadata, that can help users with a multitude of actions, automate complex processes, and improve productivity. »
Salesforce is expanding its Actions Library with new features that let sellers benefit from conversational AI directly in Sales Cloud. Instead of accessing a bespoke AI tool disconnected from their CRM, salespeople can simply open Einstein Copilot and request a personalized close plan, information on why an opportunity cannot be closed within the month , or check to see if pricing was discussed on a recent call – meaning more time spent with the customer and less time sifting through data and screens.
Einstein Copilot includes a reasoning engine that can interpret user requests and dynamically automate tasks from a library of actions, ensuring that each new action expands the range of use cases it can support – and does so without needing to train or refine a large number of cases. language model (LLM).
Einstein Copilot now includes new sales actions to increase salesperson productivity across a wide range of cross-functional use cases:
- Close packages: To accelerate closing, salespeople and managers can ask Einstein Copilot to create a personalized, grounded closing plan for an opportunity with step-by-step tactics based on opportunity history and account data, as well as recommended action dates.
- Forecasting Tips: Sales leaders can understand and reduce forecasting risks with Einstein Copilot. They can ask questions like “what transactions are at risk?” » and receive a summarized list of offers that could explain why their team member might not meet their quota.
- Call Explorer: Using recovery augmented generation (RAG), which allows companies to use their proprietary structured and unstructured data to make AI more reliable and relevant, sales reps can query previous call transcripts captured in Einstein Conversation Insights, asking questions like “what was the customer’s feeling during this call? and receive a semantically relevant response tailored to that call to avoid hallucinations.
- Follow-up emails: To help move deals forward, sales reps can ask Einstein Copilot to create personalized follow-up emails based on previous calls, which will help them move deals forward.
- Einstein Copilot in Sales Cloud Everywhere: Customers can take Einstein Copilot wherever they work. They can research leads, stay up to date with deals, and write outreach materials from any web page with Sales Cloud everywhere.
In addition to these new seller features, Einstein Copilot introduces platform enhancements that improve every conversational AI experience:
- Co-pilot analyses: Administrators can visualize Einstein Copilot usage through a pre-configured analytics dashboard, allowing them to understand and accelerate Einstein Copilot adoption by dissecting key metrics such as actions used, average interactions per user and success rates to help businesses generate and measure ROI from AI. .
- Recommended actions: Teams are now able to quickly complete any task with recommended actions in the workflow. Instead of entering a request, Einstein Copilot provides standard one-click actions specific to the page a user is on, such as “summarize opportunity” on the opportunity page or “compose email” on the contact page.
- Einstein Copilot on the Salesforce mobile app: For AI-powered productivity in the workflow, including on the go, Einstein Copilot is now available on the Salesforce mobile app, with voice-to-text functionality for mobile users that lets them simply communicate with their data and their workflows.
A unified AI assistant powered by trusted data and metadata
Einstein Copilot stands out from other AI assistants and chat solutions because it is a unified conversational AI assistant built on the Einstein 1 platform, based on business-specific data stored in Data Cloud and connected to Salesforce metadata – data that describes the unique characteristics of a customer. business setups.
Using this metadata framework, a salesperson can ask Einstein Copilot questions such as “Can you create an email for my best deal open this week?” » A typical LLM cannot find the relevant result from data alone. They need to know what agreement means and where to find the relevant data to answer the question. Unlike data, which refers to records within Salesforce (such as accounts, contact information, and sales opportunities), metadata describes the presentation, behavior, and relationships of data.
Einstein Copilot can use this metadata to interpret the user prompt in its full context, working with the LLM to choose the relevant fields (in this case, for top open offers related to a particular user), then create a personalized email for seller.
All of these interactions with Einstein Copilot are secured by the Einstein Trust Layer, which uses security and privacy measures including data masking, a zero-retention architecture to ensure data is never stored outside of Salesforce, a toxicity detection to protect responses and a client. an audit trail owned to enable AI-driven productivity without compromising trust.
What’s next: Salesforce is investing in a comprehensive library of people- and industry-specific actions that expand the range of tasks Einstein Copilot can perform. In addition to today’s previously published news on business stocks and banking stocks, as well as healthcare stocks and service stocks, such as AI-generated answers based on the knowledge base of a customer, Salesforce is developing new marketing actions that accelerate campaign creation, sales actions that help buyers, as well as the ability to transform MuleSoft APIs into custom actions that allow Einstein Copilot to distribute work across multiple systems.
Source: Salesforce
