AI Brain and trusted sources

Give AI grounded business context without allowing it to invent policy, product details, or customer outcomes.

AI Brain

Set the business identity, allowed topics, prohibited topics, tone, response rules, and handoff rules. This is grounded prompt context, not model fine-tuning.

Do not add API keys, passwords, personal data, legal/medical guarantees, unverified outcomes, or changing prices as AI policy.

Answer sequence

The assistant checks the active flow, catalog, approved trusted sources, and AI Brain policy. If no confirmed answer exists, it creates a pending agent handoff rather than guessing.

Sources can be FAQ text, document extracts, or approved links. Add only content your team has reviewed.