Salesforce launches Einstein Studio for training AI models with Data Cloud

Salesforce is moving the needle on AI. The Marc Benioff-led CRM leader today announced the launch of Einstein Studio, a new “bring-your-own-model” experience that allows enterprises to connect and train their own AI models on proprietary data within Salesforce.

The offering streamlines the AI project lifecycle, making it possible for data science and engineering teams to manage and deploy models more quickly, efficiently and at a lower cost. Once trained, these models can power any sales, service, marketing, commerce and IT application within Salesforce, the company said.

“Einstein Studio offers a faster, easier way to create and implement custom AI models … Now, Salesforce customers can harness their own proprietary data to power predictive and generative AI across every part of their organization,” Rahul Auradkar, EVP and GM for unified data services and Einstein at Salesforce, said in a statement.

The offering has already been tested by multiple enterprises as part of a pilot and is now generally available for all users of Salesforce’s Data Cloud.

How will Einstein Studio help Salesforce customers?

Every enterprise today is racing to build and deploy AI models targeting different business-critical use cases such as predicting future demand or delivering better recommendations. However, the task of building and deploying enterprise-ready AI across key applications and workflows is very exhaustive in itself. Teams have to extract, transform and load (ETL) data to prepare it for AI platforms, train the models and then implement while monitoring the entire lifecycle of the project end-to-end. This takes a lot of time and resources, making it difficult for teams to deploy their projects when needed. 

According to a KPMG survey, nearly 60% of U.S. executives say they are still a year or two away from implementing AI solutions. 

With Einstein Studio, Salesforce is making the entire process of deploying AI faster. The offering allows users to connect custom AI models built with external services such as Google Vertex AI and train them on data hosted within the Salesforce Data Cloud to solve specific business needs. 

Salesforce Data Cloud brings together data points from different sources to host unified customer profiles that adapt to each customer’s activity in real time. Einstein Studio’s pre-built, zero ETL integration leverages this data directly for model training. All the user has to do is just “point and click” on relevant data assets within the data platform.

Further, the BYOM solution also provides a control panel for managing the use of the AI models being trained, empowering data scientists and engineers to govern how their data is exposed to AI platforms for training. 

“Einstein Studio streamlines the entire AI project lifecycle – from data acquisition and prep with Data Cloud to modeling, model deployment and insights consumption. Our bring-your-own model approach helps organizations tackle their highest-value AI use cases by fully integrating the business, IT, data professionals and the end user, allowing organizations to leverage their investment in the latest AI platforms,” Sanjna Parulekar, VP of product marketing at Salesforce, told VentureBeat.

Deployment within and outside Salesforce

Once a model is trained with Data Cloud, it can be integrated into various Salesforce experiences, including Data Cloud, Flow and Apex and power company applications. For example, a propensity-to-buy model built using AWS SageMaker and registered in Einstein Studio could be used in a Flow automation to inform whether or not a product discount email should be sent to a customer.

Similarly, Parulekar explained, customers and independent software vendors will continue to have the flexibility to utilize these models in external applications. Retailers, for instance, can use the trained models to recommend products to customers based on their interests and behaviors, personalize pricing based on their individual needs or segment customers into different groups based on demographics, purchase history, etc.

“Through Salesforce’s Data Cloud and Einstein Studio, customers will now have the ability to bring their own … models, providing them greater choice in how they utilize AI and customer data. The democratization of such a capability is key to the success of our clients. Deloitte is excited to be a part of this journey and has developed a series of ‘bring your own models’ that our clients can leverage as part of the Salesforce ecosystem,” said David Geisinger, global alliance lead at Deloitte Digital.

Currently, Einstein Studio allows users to build a custom model from scratch or connect from AWS SageMaker and Google Vertex AI. However, Parulekar did confirm that the company will add more services in the future. The offering will be automatically enabled for all users of Salesforce Data Cloud starting today.