Dive Brief:
- Salesforce's generative AI assistant, Einstein Copilot, will provide responses using unstructured data within the company's Data Cloud product, the company announced Thursday, along with a slate of updates aimed at easing data constraints.
- Salesforce will pilot unstructured data support in Data Cloud in February. The feature will let Einstein Copilot scour PDFs, emails, audio, social media content and other unstructured data for answers to user queries, the company said.
- The new capability, currently available in a closed beta, will rely on the company's Einstein Trust Layer to pick accurate responses with relevant citations.
Dive Insight:
One focal point for enterprise generative AI plans is data. Model outputs hinge on the quality of each input.
Salesforce's latest feature announcement aims to ease those concerns for customers across a number of potential use cases, including customer service, call centers, marketing and sales.
With data representing a key determinant of AI success, the majority of IT leaders say they plan to increase investments in data infrastructure and AI adoption in 2024, according to an October report from the MIT Technology Review and commissioned by Databricks.
Major providers of AI technology have spent the year adding features that address enterprise pain points for adoption, especially as they relate to data privacy. In August, OpenAI released an enterprise tier of its ChatGPT service with additional data security capabilities built in. Google Cloud also embedded secure data functionalities in March as it rolled out its Generative AI App Builder in March.
Salesforce in September moved to rebrand its widely adopted CRM platform as Einstein 1, amid plans to bolster the platform's AI capabilities, and announced a hiring push across units driven by its AI expansion plans.
