Considerations for Agentforce for Energy & Utilities

Supported LLMs and Locales

Agentforce for Energy & Utilities supports select large language models (LLMs) and routes requests through the LLM gateway. It supports specific languages and locales.

The Einstein Trust Layer

The solution integrates with the Einstein Trust Layer to ensure data security and privacy.

• LLM Data Masking: The trust layer identifies and masks sensitive data, such as personally identifiable information (PII) and payment card industry (PCI) data, in prompts before sending them to the LLM. This data masking prevents sensitive data exposure to the LLM and keeps CRM data secure within Salesforce.
• Zero Data Retention: Salesforce partners with third-party LLM providers to enforce a zero-data retention policy. Under this policy, no data is used for model training, retained by third-party LLMs, or reviewed by humans.
• Toxicity Detection: The trust layer identifies and classifies toxic content. Toxic content refers to content containing violent, abusive, hateful, or sexual language, imagery, or themes that carry a risk of harm to individuals or groups.

Usage and Billing

• Consumption Tracking: You can monitor and track usage credit consumption through Digital Wallet in Setup.
• Billing Types: Depending on your agent type, customers choose between Conversations and Flex Credits for consumption billing Flex Credits are the required usage type for employee-facing agents.