One of the more prominent mottos planted across last week’s Dreamforce conference touted “everyone’s an Einstein.” First introduced in 2016, Einstein has become synonymous with “AI at Salesforce”, now integrated within all products to “discover insights, predict outcomes, increase productivity, and deliver on the promise to democratize AI” for customer companies.

To that end, announced at last week’s San Francisco event, the Einstein One Platform has Data Cloud natively integrated as a metadata framework for intelligent productivity tools for either low code/no code Salesforce users or more advanced developer options. Alongside all these products now sits “Einstein Copilot,” a conversational AI assistant that taps the world of advanced Large Language Models (LLMs) and Generative AI. Salesforce is not the first to brand an AI assistant as “Copilot,” indeed others include Microsoft 365, NICE’s Enlighten, Twilio, and even SAP from years back. Everyone has a Copilot!

Einstein Copilot allows users to use their own words to ask questions and create prompts, while proactively recommending actions after a query with Generative AI. To build and customize these applications there is Einstein Copilot Studio, which will include a number of prebuilt applications for consumer-facing channels including real-time chat, Slack, WhatsApp, or SMS. Einstein Copilot integrates into all Salesforce product lines including sales, marketing, commerce, and customer service.

Meet Your AI Assistant: Einstein Copilot
Specifically for service, Einstein Copilot could automatically respond to customers with automated, personalized, relevant answers, as well as enabling customer service agents to improve productivity by suggesting answers, summarizing calls, or seamlessly integrating into post-interaction workflows and business processes. Einstein Copilot is currently only available as an employee-facing tool (allowing agents to review Einstein’s output and modify it as necessary). Integration into customer-facing, self-service applications will be generally available in the near future.

The wait-and-see approach to Generative AI for customer conversations is based on Salesforce’s commitment to security and trust. This notion was also widely publicized at Dreamforce, emphasized in a number of keynotes and panel sessions by stating “your data is not our product.” Trust is layered throughout, literally, with the Einstein Trust Layer. Introduced a few months ago, the goal is to enhance AI capabilities by delivering secure data retrieval, detecting toxicity, data masking, a zero retention policy, and auditing services.

While as promising (and hyped!) as Generative AI is, the challenges and risks associated with Generative AI are serving as a safeguard in Salesforce’s ongoing quest to find a “single source of truth for customers.”

Categories: Conversational Intelligence, Intelligent Assistants, Articles