> ## Documentation Index > Fetch the complete documentation index at: https://ago.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Set Up Your Knowledge Base > Learn how to build, maintain, and optimize a knowledge base for your AI agents Your agents are only as good as the knowledge they draw from. This tutorial walks you through building a high-quality knowledge base — from connecting your first source to setting up ongoing maintenance. ## Prerequisites * Admin or staff access to AGO * Content to import (help center, wiki, docs, or files) *** When a user asks a question, the agent searches your knowledge base for relevant documents, then uses those documents to formulate a response. This means: * **Coverage matters** — if a topic is not in your knowledge base, the agent cannot answer it * **Quality matters** — poorly written or outdated content leads to poor answers * **Structure matters** — well-organized documents with clear titles and headings are easier for the agent to find and use AGO supports many content platforms. Pick the one that matches where your content lives: | Content Location | Connector | Best For | | ----------------- | ------------------------------------------------------ | -------------------------- | | Notion | [Notion](../data-connector/notion-integration) | Wikis, internal docs | | Confluence | [Confluence](../data-connector/confluence-integration) | Enterprise knowledge bases | | Zendesk | [Zendesk](../data-connector/zendesk-integration) | Help center articles | | Intercom | [Intercom](../data-connector/intercom-integration) | Support articles | | Files (PDF, DOCX) | [Manual Upload](../data-connector/manual-integration) | Static documents | | API | [JSON](../data-connector/json-integration) | Custom data sources | See the full list of [Knowledge Connectors](../data-connector/knowledge-connectors). You can connect multiple sources to a single agent. Start with one, validate quality, then add more. 1. Navigate to **Knowledge** > **Sources** 2. Click **Add Source** 3. Select your connector 4. Follow the authentication steps (each connector has specific requirements) 5. Select which content to sync — you can choose specific pages, spaces, or categories 6. Click **Save and Sync** The initial sync may take a few minutes depending on the volume of content. You can track progress on the Sources page. Once the sync completes: 1. Go to **Knowledge** > **Sources** and click on your source 2. Check the document count — does it match what you expected? 3. Browse a few documents to verify content was imported correctly Common issues: * **Missing documents** — check connector permissions and scope settings * **Partial content** — some connectors have content size limits * **Formatting issues** — complex layouts may not import perfectly AGO automatically scores your content for agent readability. 1. Navigate to **Knowledge** > **Quality** 2. Review the overall quality score 3. Look at individual document scores and flags | Score | Meaning | Action | | ------ | -------------------------------- | ---------------------------- | | High | Well-structured, clear, complete | No action needed | | Medium | Some issues but usable | Improve when possible | | Low | Significant problems | Fix before using with agents | Common quality issues: * **Missing titles** — documents without clear titles are hard for agents to find * **Thin content** — very short documents may not contain enough information * **Outdated information** — stale content leads to incorrect answers * **Duplicate content** — multiple documents covering the same topic confuse retrieval See [Knowledge Quality](../features/knowledge-quality) for a full reference of quality dimensions. For each flagged document, decide whether to: 1. **Fix at the source** — update the document in Notion, Confluence, etc., then re-sync 2. **Use change proposals** — suggest edits directly in AGO for team review 3. **Exclude the document** — remove low-quality content from the agent's scope Focus on fixing the documents that agents use most frequently. Check analytics to see which documents appear in the most conversations. See [Knowledge Change Proposals](../features/knowledge-change-proposals) for the review workflow. If your users speak multiple languages: 1. Navigate to **Knowledge** > **Translations** 2. Select the languages you want to support 3. AGO can auto-translate your content or you can provide manual translations 4. Review translated content for accuracy See [Knowledge Translations](../features/knowledge-translations). A knowledge base requires regular updates. Set up these practices: | Frequency | Task | | ------------- | -------------------------------------------------------------------------------------------- | | **Automatic** | Connector sync keeps content up to date | | **Weekly** | Review quality scores, fix flagged issues | | **Monthly** | Audit for outdated or missing content | | **As needed** | Use [update requests](../features/knowledge-update-requests) to flag content needing changes | #### Enable Versioning Track changes to your knowledge base over time: 1. Navigate to **Knowledge** > **Versioning** 2. Enable version history 3. Review changes before they go live See [Knowledge Versioning](../features/knowledge-versioning). ## Checklist * [ ] Connector chosen and source connected * [ ] Initial sync completed successfully * [ ] Document count verified * [ ] Quality analysis reviewed * [ ] Critical quality issues fixed * [ ] Translations added (if needed) * [ ] Maintenance schedule established *** ## What's Next * [Knowledge Best Practices](./knowledge-best-practices) — strategies for high-quality content * [Connect Notion as a Knowledge Source](./connect-notion-howto) — detailed Notion setup * [Knowledge Overview](../features/knowledge-overview) — full feature reference * [Build Your First Agent](./getting-started-tutorial) — use your knowledge base with an agent