Salesforce goes one step further with its AI assistant. Einstein Copilot examines your data closely to understand everything better.

The Dreamforce event is the ideal opportunity for Salesforce to introduce the next generation of Einstein. Salesforce is putting the name of the legendary German physicist on a wide range of AI features within its platform. In March, Einstein made the leap from analytical to generative AI with Einstein GPT. The next step in this development is Einstein Copilot and a supporting data platform Einstein 1.

First, let’s look at Einstein 1. This platform must become the connective tissue between the two spearheads of the Salesforce platform: Data Cloud and Einstein, but also the entire portfolio. Einstein provides the ability to connect data from the data cloud to build AI-driven low-code applications. Every moment of interaction with a customer can be optimized based on the data you have about your customer.

Einstein co-pilot

The eye-catcher on Dreamforce’s opening day, however, is Einstein Copilot. Salesforce introduced EinsteinGPT across its portfolio in March and is now taking it a step further. Einstein Copilot can not only answer questions about your data, but also perform proactive analysis and suggest actions for you. The AI ​​assistant can be used in different departments and for different purposes.

• Sales: Prepare and automatically summarize customer conversations. After the conversation, Einstein Copilot suggests follow-up actions. With Copilot you can also generate customer emails or clauses for sales contracts.
• Service: Optimize customer service based on real-time data. Summarize complaints and propose solutions through the customer’s chosen communication channels.
• Marketing: Create email campaigns and identify the right target groups using Data Cloud data. Einstein Copilot also creates landing pages based on each customer’s personalized preferences.
• E-commerce: Leave the management of your web shop to Einstein Copilot. Copilot supports you step by step in setting up your web shop to increase conversions. Product descriptions can be generated in multiple languages ​​and Copilot also takes SEO into account.
• Development: Einstein Copilot can convert natural language instructions from the development environment into code, but also makes proactive suggestions for the accuracy and security of your code.
• Tableau: Einstein Copilot transforms raw data in Tableau into more actionable insights based on your questions to the Chabot.

Salesforce emphasizes that Einstein Copilot can also be used for industry-specific applications. The pilot version is available now.

In the studio

Einstein Copilot is also accompanied by a Copilot Studio. This follows Einstein Studio, which Salesforce launched in August. Using the Einstein (Copilot) Studio, you can give generative AI models access to your own data in the Data Cloud. This also allows you to train these models on specific company data, making the output more relevant.

The Einstein Copilot consists of three parts. The first is a Prompt Builder with which you can teach Einstein your company’s corporate identity. Einstein Copilot generates prompts based on customer data it pulls from the Data Cloud, in a style that suits your business. This can be useful for generating marketing messages, but you can also have Copilot communicate directly with your customers via the website, Slack, WhatsApp or SMS.

With the Skill Builders Develop specific tasks for Einstein Copilot based on data from the Data Cloud. If you have already created models on an external data platform, you can do this through the model maker can connect to Einstein Copilot Studio to further train the model with Data Cloud data. These integrations are offered for Databricks and Snowflake, among others. The Einstein Copilot Studio will be open from fall 2023.

With trust

Finally there is that layer of trust within the Einstein Copilot platform. Salesforce recently found that seven in 10 consumers are concerned about how companies are handling AI. The trust layer must ensure that data from companies and individuals is treated with respect for the rules and also prevents Einstein from providing answers that offend users. This security layer is built into every Einstein product by default.