Salesforce describes 'grounding with your data' as part of the Einstein Trust Layer's value proposition. In the context of Data 360 retrievers, what does this phrase most directly refer to?
Select an answer to reveal the explanation.
Short Explanation
'Grounding with your data' means the agent leans on Ridgeline's own indexed documents and records for its answers, instead of just guessing from whatever the model happened to learn during training.
Full Explanation
'Grounding with your data' captures the idea that an agent's responses can be based on an organization's own content — its indexed documents in the Data Library, its CRM records, and so on — rather than relying solely on the LLM's generic pretrained knowledge, which has no awareness of Ridgeline's specific policies, patients, or documents. This is precisely why retrievers and record-based grounding both matter: they are the mechanisms that make 'your data' actually reach the model's context. It does not mean customer data is never processed by a language model at all — grounding necessarily involves relevant data being included in the prompt sent to the model, with the Trust Layer's separate protections (like masking and zero-retention) managing how that data is handled, not preventing its use entirely. It also doesn't restrict grounding to structured CRM records only — the whole point of the Data Library and retrievers is to extend 'your data' grounding to unstructured documents too, which this claim would wrongly exclude. Automatic post-conversation data deletion describes something closer to a zero-retention policy, a related but distinct Trust Layer concept, not what 'grounding with your data' itself refers to. For the exam, keep this phrase tied to using an org's own content as the basis for responses.