Salesforce Inc. today announced a set of integrated tools that data scientists and engineers can use to build secure artificial intelligence applications using their own company data in the Salesforce Data Cloud.

Einstein Studio uses a “bring your own model” approach that works with Amazon Web Services Inc.’s SageMaker, Google LLC’s Vertex and other popular services used to build and train AI models.

Salesforce describes Einstein Studio as the “first real-time data platform for customer relationship management” and said it can be used across a wide variety of sales, service, marketing, commerce and general information technology applications within Salesforce. It cited a recent KPMG LLC survey that found that 60% of U.S. executives say they are still at least a year away from implementing their first generative AI applications.

“Most customers struggle with the ‘last mile’ of an AI project, getting outputs, predictions and insights into the hands of business users,” Rahul Auradkar, executive vice president and general manager of unified data services and Einstein, said in e-mailed comments. “Einstein Studio solves this challenge with native integration with Data Cloud and the Salesforce Platform. Model inferences are available in real-time in workflows ensuring AI projects drive better business outcomes.”

Organizations can use Einstein Studio with real-time customer data from the Data Cloud, which connects customer data from any source and aggregates it into a single customer profile. The platform eliminates the need for an extract/transfer/load stage when used with SageMaker and provides point-and-click access to data in the data cloud. The company has also published a Python software development kit for connections to other sources.

A control panel enables data scientists and engineers to govern how their data is exposed to chosen AI platforms for training. Einstein Studio uses built-in Data Cloud functions for data modeling, data harmonization, unification and various data quality-related transforms, Auradkar said.

“Users can turn AI models into actionable insights, helping guide Flow automation, trigger Apex code or inform sales reps and contact center agents with AI outputs surfaced across the Lightning experience,” he said, referring to Salesforce’s workflow manager, object-oriented programming language and component-based development framework. “The integration with Data Cloud allows for use of the model outputs for activation into customers’ choice of apps and channels through the rich programmatic interfaces that Data Cloud supports.”

The BYOM approach “helps organizations tackle their highest value AI use cases and leverage their investment in the latest AI platforms,” Auradkar said. Einstein Studio comes as part of Data Cloud and is priced based on consumption.

Image: Salesforce