Salesforce Debuts ‘Bring-Your-Own-Model’ Generative AI Platform Einstein Studio
Salesforce has expanded its generative AI toolkit for clients with the new “bring-your-own-model” Einstein Studio platform. Companies can import and train their AI models on proprietary data within the Salesforce platform to streamline the creation and deployment of the models for enterprise tasks in marketing and sales.
Einstein Studio is designed for companies that have already begun building AI models with Amazon SageMaker and other locations. Integrating those models into other locations is often difficult and expensive, but Einstein Studio pitches itself as a no-code alternative that brings the data produced by those models into the Salesforce Data Cloud. The service is starting with the option of linking to Amazon SageMaker but has plans to incorporate Google Vertex AI, Databricks, and other services.
Enterprises can use Einsteing Studio to combine their custom AI models with Salesforce’s Einsteing GPT large language model to carry out tasks like automated maintenance reminders and marketing emails. At the same time, they’ll be able to employ the models in other contexts, a flexibility that Salesforce claims is in high demand among its customer base.
“Companies need quick, ROI-driven AI investments that deliver value through actionable business insights and personalized customer experiences,” Salesforce executive vice president and general manager of unified data services and Einstein Rahul Auradkar said. “Einstein Studio offers a faster, easier way to create and implement custom AI models, including a BYOM approach that allows customers to use the most relevant AI models – all while bypassing expensive ETL data pipeline processes. Now, Salesforce customers can harness their own proprietary data to power predictive and generative AI across every part of their organization.”
Einstein All Over
Einstein Studio is Salesforce’s latest addition to its AI Cloud portfolio of enterprise generative AI tools. Einstein’s resurrection earlier this year has led to Salesforce hosting a range of models and investing $500 million in related startups. It also supports models built in-house by Salesforce for business automation and writing code. Businesses can also incorporate their own custom models into the platform.
It all passes through the Einstein Trust Layer, a moderation and control system that monitors and works to maintain business and regulatory compliance rules by stopping the models from holding onto any private or sensitive information. The Einstein Trust Layer watches for when sensitive data might be sent by a user on an app and stops it from getting to the model, alerting the user that they might not want to submit that information. Einstein also watches for and filters out toxic and discriminatory language coming from both ends.