Knowledge base
The collection of business documents an AI agent retrieves from — policies, product information, FAQs — so its answers reflect your business rather than general internet knowledge.
One term from the WhatsApp Business Platform glossary. Last updated 2026-07-25.
The knowledge base is what makes retrieval-augmented generation useful. Return policies, shipping timelines, product specifications and pricing rules become retrievable context the model sees before answering.
Quality here is upstream of everything else. An agent grounded in an outdated policy document will answer confidently and wrongly, which is why knowledge base freshness matters more than model choice for most support deployments.
Related terms
- Retrieval-augmented generation (RAG)
- A technique where an AI agent retrieves passages from your own knowledge base before answering, so replies are grounded in your documented policies and product facts rather than invented.
- Hallucination
- An AI model producing a confident but unsupported answer. In customer messaging it looks like an invented policy, price or delivery promise.
- Tool calling
- Letting an AI agent invoke a real function — looking up an order, checking availability, calling your API — instead of only generating text.
See Knowledge base in a real workspace
BotPulsar runs the official WhatsApp Business Platform with unlimited team members on every plan, starting at ₹300/month.