Skip to content

Salesforce simplifies access to external data sources for AI model training

Eyeing the need for organizations to train their own artificial intelligence models, Salesforce Inc. today is announcing a new feature of its Data Cloud data lake that collects information from a variety of third-party sources.

Called Unified Knowledge, the feature can be used to gather and curate data needed to train AI models for use by customer service agents. Citing Accenture Plc research that found that three-quarters of executives struggle to scale AI projects because of disconnected systems and siloed data, Salesforce said Unified Knowledge can be used to import unstructured data into the Data Cloud for transformation, tagging and quality assurance.

Data is primarily intended to be used with the company’s Einstein for Service customer service support application, but the data integration feature can also be applied to the company’s Sales Cloud, Health Cloud, Financial Services Cloud and Field Service applications. The capability feature was developed through a partnership with Zoomin Inc.

While intended mainly for use with Data Cloud and the Einstein Trust Layer governance platform, customers can also use third-party large language models if they choose. “We have partnerships with several generative AI partners,” said Kishan Chetan, general manager of Salesforce Service Cloud. “We’ve done this as bring-your-own-model.”

With Unified Knowledge, AI models are informed by Salesforce knowledge articles and resources from third-party sources such as Microsoft Corp.’s SharePoint, Atlassian Corp. PLC’s Confluence, Google LLC’s Drive and YouTube, Amazon Web Services Inc.’s S3 storage, Adobe Systems Inc.’s Experience Platform, Guru Technologies Inc.’s Guru, Zendesk Inc.’s customer service platform and company websites.

Chetan described the administrative setup process as “relatively simple. Within our knowledge management tool, we have tagging tools, but once you put the content into gen AI, a lot of the content can be automatically determined,” he said.

Data is stored as Salesforce Knowledge Objects, and users can use the company’s Prompt Builder to create generative AI prompts. “As we evolve, we will be able to give you extra knowledge inferences. Right now, it is married to data already in Data Cloud,” Chetan said.

Data from external sources can be used to train customer-facing answer bots, streamline employees’ access to internal information and quickly search company knowledge bases.

Access to data in external applications is provided through published application program interfaces and access controls are governed by the Salesforce access layer.

Unified Knowledge is available in a free beta test for Salesforce customers with Service Cloud Unlimited Edition, Einstein 1 Service Edition or the Knowledge Add-On. A freemium version of Unified Knowledge will continue to be included with those applications. Salesforce Lightning Knowledge is required. Classic Knowledge is not supported.

Knowledge customers with an authoring license will have access to the freemium version, regardless of whether they got Knowledge licenses as part of Service Cloud or purchased them for another edition. Customers must negotiate with Zoomin on custom quotes for connectors.

Photo: Salesforce

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU

Leave a Reply

Your email address will not be published. Required fields are marked *