Salesforce adds Einstein Copilot Search, vector database to its Data Cloud, to help enterprises take advantage of unstructured data for analysis, and build AI-based applications. Credit: SARINYAPINNGAM Salesforce is updating its Data Cloud with vector database and Einstein Copilot Search capabilities in an effort to help enterprises use unstructured data for analysis. The customer relationship management (CRM) software provider’s Data Cloud, which is a part of the company’s Einstein 1 platform, is targeted at helping enterprises consolidate and align customer data. The Einstein 1 platform, in turn, is a data engine with a low code and no code interface that is designed to let enterprises connect data to build AI-based applications. As part of the updates, Salesforce has integrated vector database support via the Data Cloud Vector Database feature, which makes it easier for the Data Cloud to handle diverse data types. “This database allows Salesforce customers to combine structured and unstructured data, creating a more comprehensive customer profile,” the company said in a press release, adding that once the unstructured data is added to the Data Cloud, it is automatically converted into a usable format across the Einstein 1 platform. This, according to the company, makes the unstructured data available for analysis and utilization across various workflows within Salesforce applications including Flow, Apex, and Tableau. Salesforce has also added an AI search capability to Einstein Copilot, which will allow the generative AI-based assistant to interpret and respond to complex queries from enterprise users by tapping into diverse data sources, including unstructured data. “Copilot Search will provide precise, contextually relevant responses in a user’s workflow and bolster trust with source citations from the Einstein Trust Layer,” the company said. The Einstein Trust Layer is based on a large language model (LLM) built into the platform to ensure data security and privacy. In order to take advantage of unstructured data via Einstein Copilot Search, enterprises would have to create a new data pipeline that can be ingested by the Data Cloud and stored as unstructured data model objects. These data model objects have to be transformed into data fit for use in AI applications by converting the data into embeddings, which are numeric representations of data optimized for use in AI algorithms, the company said, adding that these embeddings are then indexed for use in search across the Einstein 1 platform alongside any other existing structured data. The Einstein Copilot Search capability can also be paired with retrieval augmented generation (RAG) tools — which Salesforce supplies — in order to enable Einstein Copilot to answer customer questions. Answers comes with semantically relevant information, citing the knowledge sources used to craft the answers, the company said. Related content news UK government’s Pensions Dashboards Programme delayed Skills shortages, inadequate governance, and rising costs stymie the PDP, but the National Audit Office says it believes progress is being made. By John Dunn May 09, 2024 4 mins Government IT Government IT Skills opinion Raj Polanki’s five steps by which CIOs can lead holistically The US Division CIO of Wacker Chemie says tech chiefs should think beyond run, grow, and transform, and consider how they are uniquely positioned to promote social values across the business and beyond. By Michael Bertha and Duke Dyksterhouse May 09, 2024 10 mins CIO Diversity and Inclusion IT Leadership brandpost Sponsored by Rocket Software How to successfully integrate data in a hybrid environment To successfully integrate data in a hybrid cloud environment, organizations must create a simple, secure, and powerful approach with the right modernization tools. By Phil Buckellew May 09, 2024 4 mins Digital Transformation brandpost Sponsored by Rocket Software Rethinking DevOps and automation with a layered approach For all its benefits, automation is not something that can just be implemented blindly across the layers of the DevOps stack. If those functions aren’t working together, the automation in each layer only adds more complication, creating ineffic By Phil Buckellew May 09, 2024 4 mins Digital Transformation PODCASTS VIDEOS RESOURCES EVENTS SUBSCRIBE TO OUR NEWSLETTER From our editors straight to your inbox Get started by entering your email address below. Please enter a valid email address Subscribe