Skip to main content

Overview

The Knowledge Base is your organization’s shared, searchable memory. It combines semantic search (matching by meaning) with keyword search, so a question phrased differently from the original text still finds the right answer. It’s what lets the NextKS bot and the Voice assistant answer “how do we…?” and “what did we decide…?” without bothering a teammate. You don’t build the knowledge base by hand — it fills up automatically as your team works.

What’s in the knowledge base

Content reaches the knowledge base through two paths:

Accepted Q&A answers

When a ticket author accepts an answer, that question-and-answer is added to the knowledge base, discoverable across your organization.

Published meeting minutes

When a meeting owner publishes a meeting’s minutes, those minutes join the knowledge base.
Raw meeting transcripts are never added to the knowledge base — only the generated minutes.

Visibility

Knowledge-base content is either private or public:
  • Private — your own content that hasn’t been shared yet (for example, meeting minutes you’ve saved but not published). Only you can find it in search.
  • Public — content that’s been published (accepted Q&A answers, published minutes). Everyone in your organization can find it.
Search always returns public content plus your own private content — never anyone else’s private content. Publishing is simply the moment your content flips from private to organization-wide; unpublishing flips it back.

Searching the knowledge base

The knowledge base is searched automatically on your behalf:
  • The NextKS bot (Slack/Teams) and the Voice assistant search it to ground their answers — reach for the knowledge base before answering from scratch.
  • AI agents connected through the NextKS MCP search it with the gateway_search_kb tool, which accepts one or more queries and returns the best-matching public and own-private content.
Because search is by meaning as well as by keyword, you don’t need to remember the exact wording — ask in your own words and the knowledge base finds the closest matches.