What the Knowledge Base does

The Knowledge Base is the business information your AI Agent can use when responding to customers. Dealism can generate knowledge from conversations, and you can add more information by file or link. Entries are displayed with categories such as faq, policy, pricing, product_info, strategy, and usecase.

How knowledge affects answers

Knowledge determines what your Agent knows and what it can say to customers. A specific, current entry gives the Agent a reliable fact to use. A vague, outdated, or conflicting entry can lead to incomplete or inconsistent answers.

Knowledge should describe your real business rules. It should not contain instructions that encourage the Agent to guess missing prices, availability, policies, or customer details.

Knowledge Base best practices

• Use one source of truth: avoid keeping two entries with different prices or policies.

• Write for a real customer question: use plain language and put the direct answer first.

• Separate different facts: keep products, policies, FAQs, and use cases in focused entries instead of one large mixed document.

• Add conditions: state country, currency, customer type, plan, date, or exception when a rule is not universal.

• Keep sensitive information out: do not upload passwords, private customer data, payment-card details, or secrets.

• Review regularly: update knowledge whenever pricing, availability, services, or policies change.

Test and maintain your knowledge

1. Ask the Agent several real customer questions, including an exception or follow-up question.

2. Confirm that names, prices, links, policies, and next steps match the source.

3. Edit unclear entries and delete obsolete or conflicting information.

4. Retest after every important update.

If the Agent does not use a new or updated entry, see Agent is not using new knowledge.