SleekFlow AI Global Knowledge Base is the central place where you can review and manage knowledge sources used across AI agents and other SleekFlow AI features.

Each AI agent has its own agent-level Knowledge Base. Sources added to an agent’s Knowledge Base are processed by SleekFlow and used to generate articles that the agent can reference during conversations.

The Global Knowledge Base gives you a workspace-level view of these sources, so you can:

• Review sources used across AI agents
• Check processing and training status
• Manage uploaded files, website sources, web search sources, and custom answers
• Replace or delete outdated sources
• Understand what knowledge has been added to different agent-level Knowledge Bases

How Global Knowledge Base works with agent-level Knowledge Base

Each AI agent has its own Knowledge Base that defines what knowledge it can use. The Global Knowledge Base provides a centralized view to review and manage these sources across your workspace.

Learn more about how to add data source to Agent-level knowledge base in our Help Center article.

Understand how sources are processed

After a source is added, SleekFlow processes and indexes its content. Depending on the source and AI feature, the processed knowledge may be used to generate articles that AI agents can reference.

Generated articles are structured pieces of knowledge created by SleekFlow based on your sources. Instead of referencing raw files or webpages directly, the AI uses these articles as the primary format for retrieving and delivering information.

Each generated article:

• Organizes content from one or more sources
• Breaks down information into clearer, AI-readable sections
• Helps standardize how knowledge is used across conversations

This makes it easier for you to review what knowledge has been ingested, and improves how the AI retrieves and responds with relevant information.

Source processing happens asynchronously. If a source has been added but the generated article is not ready yet, the assigned agent may still reference the source material directly. Once the article is generated, the agent will reference the generated article instead.

Access the data sources added to your SleekFlow workspace

You can follow the steps below to access the data sources you’ve added to your SleekFlow workspace:

1. Click on the  icon on the left navigation bar
2. Click on “Data and tools” at the top menu
   
3. In this page, you will find:
   
   1. All data sources list
       View all knowledge sources added to your workspace in one place. Each row shows:
      1. Source name: The name and URL or file reference
      2. Type: Source type (for example, Web link, File, or Custom answer)
      3. Agents: Which AI agents the source is assigned to
      4. Status: Current processing status (for example, Active)
      5. Last updated: When the source was last modified
   2. Search and filter
       Use the search bar to find a specific source by name.
       Use the “All types” dropdown to filter sources by type.
   3. Agent assignment column
       Assign or update which AI agents can use a source directly from the table.
       Click the dropdown under “Agents” to select one or more agents.
   4. Add knowledge button
       Click “+ Add knowledge” to add a new source to your workspace.
   5. Sidebar navigation
      1. All data sources: View all sources across your workspace
      2. Agents list: View sources filtered by a specific AI agent

Review generated articles

Generated articles show how SleekFlow has interpreted and structured knowledge from your sources.

Generated articles show how SleekFlow has interpreted and structured knowledge from the sources assigned to an AI agent.

To generate articles, you must first assign data sources to the agent. You can assign a source using the dropdown in the Agents column on the data sources table. Once assigned, SleekFlow processes the source and generates articles automatically.

To review generated articles:

1. Go to “Data and tools”
2. Under “Agents” section on the left-sided panel, select the AI agent you want to review.
3. Click the “Article wiki” tab.
4. Review the generated articles once they are ready.

If article generation is still in progress, you will see an update panel showing:

• The number of documents being processed
• The number of queued documents
• Each queued document name
• The current generation progress

If no articles have been generated yet, the page will show No generated articles yet.

Suggest edits to generated articles

If a generated article is inaccurate, unclear, incomplete, or outdated, you can suggest an edit.

Use suggested edits to flag knowledge that needs improvement without directly changing the generated article. Suggested edits are reviewed against the generated article’s context. If the suggestion is unrelated to the article or does not apply to its source content, it will not be accepted.

To suggest an edit:

1. Go to “Data and tools”
2. Under “Agents” section on the left-sided panel, select the AI agent you want to review.
3. Go to the agent’s Article wiki tab.
4. Open the generated article you want to review.
5. Click “Suggest edit”.
   
6. Enter your suggested changes.
   
7. Submit the suggestion.

Review article history

Article history helps you understand how generated knowledge has changed over time.

To review article history:

1. Go to “Data and tools”
2. Under “Agents” section on the left-sided panel, select the AI agent you want to review.
3. Go to the agent’s “Article wiki” tab.
4. Open the generated article.
5. Click “Version history”.
   
6. Review the available change history.