Zero Hallucination Rate
Guarantees that sensitive pricing, refund policies, legal disclaimers, or compliance guidelines are delivered verbatim.
While the semantic RAG engine synthesizes answers from long-form documents and websites, Q&A Pairs act as your ultimate single source of truth for critical business inquiries.
When a visitor asks a question matching an existing Q&A pair, the agent prioritizes your exact verified answer.
Zero Hallucination Rate
Guarantees that sensitive pricing, refund policies, legal disclaimers, or compliance guidelines are delivered verbatim.
Semantic Intent Matching
Visitors do not need to type the exact words. If the pair is “What are your delivery fees?”, the system matches “How much is shipping to Austin?” or “Do I have to pay postal charges?”.
Navigate to Knowledge Base → Q&A
Click the + New Question & Answer button.
Input the Question
Write a representative inquiry as a real user would ask it (e.g. How do I initiate a product return?).
Provide the Authoritative Answer
Compose the precise, professional response including all relevant instructions and links.
Add Question Variations (Optional)
Although our embedding model understands semantic synonyms out-of-the-box, you may append alternative phrasings for heightened certainty.
Save and Deploy
Click Save. The pair is immediately active in your agent’s knowledge index.
If you have an existing knowledge base from a legacy helpdesk (Zendesk, Intercom, Freshdesk) or an internal spreadsheet, import them in bulk via CSV:
Question,Answer"What payment methods do you accept?","We accept all major credit cards (Visa, Mastercard, Amex), Apple Pay, and ACH wire transfers.""What is your standard hardware warranty?","All hardware units come with a 2-year manufacturer warranty and optional enterprise on-site replacement.""Where is your corporate headquarters?","We are located at 500 Howard Street, Suite 400, San Francisco, CA. Support hours are 9am-6pm PT.".csv file.