Salesforce is enhancing its AI capabilities with new features and support.
Just before World Tour NYC, the company announced the integration of vector database support and improvements to its Einstein Copilot generative assistant, which now includes AI search capabilities.
Introducing AI Agents
These features will be integrated into the Einstein 1 Platform, which consolidates various components—from the Data Cloud and Copilot to CRM applications and the Copilot Studio for developing AI-driven applications. This integration aims to simplify the use of AI in team workflows, whether for creating custom generative AI apps or utilizing the Copilot assistant for project information.
Vector Database Support in Salesforce Data Cloud
Salesforce Data Cloud is the company’s proprietary data platform that synthesizes data from various customer interactions to build real-time unified customer profiles. Previously, teams had to spend considerable time and resources fine-tuning large language models (LLMs) on both structured and unstructured data to leverage generative AI in their CRM workflows.
With the new vector database support in Data Cloud, Salesforce removes the complexities associated with fine-tuning. This functionality converts unstructured business data—such as PDFs, emails, documents, and transcripts—into a vector format. By merging this data with structured information (like purchase histories and support tickets), organizations can enable generative AI and analytics across different Salesforce CRM applications. This results in robust data insights and streamlined processes, easily accessed via relevant prompts.
For example, when crafting an email campaign, a marketer can use Marketing Cloud Intelligence to analyze unstructured survey data and transcripts stored in Data Cloud to gauge customer intent. They can then refine their email templates based on these insights.
Salesforce’s emphasis on vector database support reflects the growing recognition among organizations of the vast potential held in unstructured data, which comprises 90% of total enterprise data. Competitors like Snowflake and Databricks also support both structured and unstructured data today.
“By integrating structured and unstructured data, our Data Cloud Vector Database transforms business data into valuable insights. This advancement, paired with the power of LLMs, is revolutionary, fostering a data-driven environment where AI, CRM, automation, Einstein Copilot, and analytics convert data into actionable intelligence and drive innovation,” said Rahul Auradkar, EVP and GM of Unified Data Services at Salesforce Einstein.
Einstein Copilot Enhancements
In addition to CRM applications for sales and marketing professionals, Salesforce’s generative AI assistant, Einstein Copilot, will also benefit from the new vector database functionalities within Data Cloud. This technology will empower AI-driven Einstein Copilot Search, allowing users to query all their business data—both structured and unstructured—and retrieve precise information seamlessly during their workflows.
For instance, a sales representative preparing for a customer meeting could prompt Einstein Copilot to access specific details, such as the customer’s financial filings or prior email exchanges, also asking follow-up questions to gain a deeper understanding. Salesforce anticipates that users will enjoy a similar experience across other CRM applications that utilize Einstein Copilot as a conversational aide for various tasks. Moreover, Einstein Copilot will provide citations for sourced information, ensuring users can utilize generated data with confidence.
Availability in 2024
While these advancements accelerate Salesforce’s AI initiatives, it's important to note that these features are not yet available. The company plans to launch Einstein Copilot for general use in February 2024. During this time, it will also pilot the vector database support and AI search capabilities, although a schedule for general availability has not been confirmed.
Ranking
