Salesforce AI Specialist Practice Exam

Which feature detects and replaces sensitive data with placeholder text before it is sent to the LLM?

1. Prompt Defense

2. Data Masking

3. Call Summaries

4. Work Summaries

The correct answer is: Data Masking

Data Masking is the feature responsible for detecting and replacing sensitive data with placeholder text prior to being sent to a Large Language Model (LLM). In the context of data privacy and security, this approach ensures that sensitive information, such as personally identifiable information (PII) or confidential details, is not exposed or compromised in the process of utilizing AI technologies. When data is masked, the original data is substituted with a non-sensitive version, allowing applications and users to continue functioning without directly interacting with sensitive information. This is crucial in environments where compliance with data protection regulations is necessary, as it helps organizations mitigate risks associated with data breaches. Other options do not serve this specific purpose. For instance, Prompt Defense focuses on safeguarding the prompts given to an AI model, but it does not address the replacement of sensitive data. Call Summaries and Work Summaries are more about generating overviews or summaries of interactions and work processes, rather than masking sensitive information. Therefore, Data Masking is clearly distinguished by its function of protecting sensitive data while still allowing the use of AI services.