Salesforce is enhancing its artificial intelligence products with new support and features. Shortly before the World Tour NYC in New York, the company announced that it is adding vector database support and enhancing its Einstein Copilot assistant with AI search capabilities. These features will be integrated into the Einstein 1 platform, which brings together all elements from Data Cloud and Copilot to CRM applications and Copilot Studio for building AI-driven applications. The company states that this will make it easier for teams to leverage AI in their workflows, whether it's building custom generative AI applications or calling on the Copilot assistant for the information needed for projects. Adding Vector Database Support in Salesforce Data Cloud Salesforce Data Cloud is the company's internal data platform that combines data points from various customer touchpoints to host real-time unified customer profiles. To enable generative AI in CRM workflows using this data, teams had to fine-tune pre-trained large language models (LLMs) on structured and unstructured data, which consumed a significant amount of time and resources. By integrating vector database support into Data Cloud, Salesforce eliminates the hassle of fine-tuning. Essentially, it transforms all unstructured business data, from PDFs and emails to documents and transcripts, into usable vector formats. It unifies them with structured information, such as purchase history or support tickets, to enable generative AI and analysis in various Salesforce CRM application workflows. This ultimately provides users with powerful data insights and simplified processes accessed through relevant prompts. For example, when building email campaigns, marketers can use Marketing Cloud Intelligence to analyze unstructured survey data and transcripts hosted in Data Cloud to understand the intent of their target audience. With these insights, they can iterate on email templates. Salesforce focuses on providing vector database support because most organizations recognize the potential value of unstructured data, which accounts for 90% of enterprise data. This is especially true for data science and AI. Major data ecosystem providers like Snowflake and Databricks also support structured and unstructured data today. "Our new Data Cloud vector database transforms all business data into valuable insights by integrating structured and unstructured data. This advancement in Data Cloud, coupled with the capabilities of LLM, is a game-changer that fosters a data-driven ecosystem where AI, CRM, automation, Einstein Copilot, and analytics transform data into actionable intelligence, driving innovation," said Rahul Auradkar, EVP and GM of Einstein Unified Data Services at Salesforce. Einstein Copilot to Gain AI Search Capabilities Alongside CRM applications targeted at sales and marketing professionals, Salesforce's generative AI assistant, Einstein Copilot, will also benefit from the vector database capabilities in Data Cloud. This technology will power the AI-driven Einstein Copilot search feature, allowing users of the assistant to query all their business data, whether structured or unstructured, and pull out precise information within their workflows. Imagine a sales representative preparing for a client meeting, prompting Einstein Copilot to pull out specific information such as the client's financial documents or past email interactions and asking follow-up questions for detailed insights. Salesforce states that users can expect a similar experience in other CRM applications that support Einstein Copilot as a conversational assistant for different tasks. Additionally, it will provide citations for its source material, ensuring that teams can trust the information generated by it. Available in 2024 While these features advance Salesforce's AI agenda, it is important to note that they are not currently available. The company will enable Einstein Copilot for general use in February 2024. As the conversational assistant serves customers, the company will pilot vector database support and AI search. The exact timeline for widespread availability is still unclear at this stage, although it has been confirmed that it will also happen in the same month.