Knowledge

The Knowledge Base contains your organization's documentation, indexed for AI-assisted search. When you interact with Hublvu, the AI automatically references relevant knowledge to provide contextual, accurate responses.

Why Knowledge Matters

The AI assistant is more effective when it has access to your organization's specific information:

• Contextual answers — Responses informed by your procedures and systems
• Accurate guidance — References your actual documentation, not generic advice
• Institutional memory — Captures expertise that would otherwise be lost

Knowledge Types

Institutional Knowledge

Operational documentation created by your organization:

• Runbooks — Standard operating procedures
• Playbooks — Incident response procedures
• Architecture docs — System design and topology
• Troubleshooting guides — Known issues and resolutions
• Process documentation — How things work internally

Product Files

External documentation from vendors and products:

• Product manuals — Vendor documentation
• Technical specifications — API docs, configuration references
• Release notes — Version-specific information
• Best practices — Vendor recommendations

Uploading Knowledge

Accessing Knowledge Management

Administrators can access Knowledge Management:

1. Expand the Knowledge section in the sidebar
2. Click Institutional Knowledge or Product Knowledge

Uploading Files

1. Click Upload
2. Select files from your computer
3. Add metadata (optional but recommended):
   • Title
   • Description
   • Labels/tags
4. Click Upload

Supported Formats

Format	Extensions
Documents	PDF, DOCX, DOC
Spreadsheets	XLSX, CSV
Text	TXT, MD

Processing

After upload, files go through processing:

1. Parsing — Content extracted from file
2. Chunking — Content split into searchable segments
3. Embedding — Semantic indexing for AI search
4. Indexing — Made available for queries

Processing status is visible in the file list. Large files may take several minutes.

Managing Knowledge

Viewing Files

The knowledge list shows:

• File name
• Upload date
• Processing status
• Source/category
• Labels

File Details

Click a file to see:

• Full metadata
• Processing status
• Content preview
• Usage statistics

Organizing with Labels

Labels help categorize knowledge:

1. Create labels that match your organization (by team, system, topic)
2. Apply labels when uploading or editing
3. Filter by label when browsing

Updating Content

To update a document:

1. Upload the new version
2. The system processes it as new content
3. Optionally remove the old version

Removing Knowledge

Delete files no longer needed:

1. Select files to remove
2. Click Delete
3. Confirm deletion

Deleted files are removed from AI search immediately.

Product Files

Product files include additional metadata extraction:

Automatic Metadata

When you upload product documentation, the AI extracts:

• Company/Vendor — Who produced the document
• Product — What product it covers
• Version — What version it applies to
• Summary — Brief content overview

Filtering

Use extracted metadata to filter:

• By vendor
• By product
• By version

This helps find relevant documentation quickly.

Auto-Refresh

For knowledge sources that update regularly:

Configuring Auto-Refresh

1. Open file details
2. Enable Auto-Refresh
3. Set refresh interval
4. Provide source URL (if applicable)

The system re-fetches and re-processes content on schedule.

How AI Uses Knowledge

In Chat

When you ask questions, the AI:

1. Searches relevant knowledge
2. Retrieves matching content
3. Uses it to inform the response
4. May cite sources

Example:

• You ask: "What's our procedure for database failover?"
• AI searches knowledge for "database failover procedure"
• Finds your runbook
• Provides answer based on your documentation

In Flows

Flows can reference knowledge:

• Triage steps may search for relevant documentation
• Remediation steps may reference procedures
• The AI uses knowledge to make better decisions

In Guides

Guide creation can pull from knowledge:

• AI generates steps based on existing procedures
• References relevant documentation
• Ensures consistency with established practices

Searching Knowledge

From Chat

Simply ask:

• "Find documentation about authentication"
• "What do we have on the payment system?"
• "Search for Kubernetes deployment procedures"

Direct Search

In the Knowledge section:

1. Use the search bar
2. Enter keywords or phrases
3. Review results
4. Click to open full documents

Semantic Search

Hublvu uses semantic search, meaning:

• It understands meaning, not just keywords
• "Database connection issues" finds content about "DB connectivity problems"
• More flexible than exact keyword matching

Best Practices

Keep Knowledge Current

Outdated documentation leads to outdated AI responses. Update or remove stale content.

Use Descriptive Names

Clear file names help both humans and AI find the right content.

Add Labels Consistently

A consistent labeling scheme makes knowledge easier to organize and find.

Upload Complete Documents

The more context in a document, the better the AI can use it. Avoid uploading fragments.

Review AI Responses

When the AI references your knowledge, verify it's using current information. This helps identify outdated content.

Troubleshooting

File Not Appearing

If an uploaded file isn't available:

• Check processing status (may still be processing)
• Verify file format is supported
• Check for processing errors

Poor Search Results

If search isn't finding relevant content:

• Try different search terms
• Check if the content has been uploaded
• Verify processing completed successfully

AI Not Using Knowledge

If the AI doesn't reference your documentation:

• Ensure relevant content exists
• Try more specific questions
• Verify the knowledge is processed and indexed

Related Topics

• Chat — Ask questions using knowledge
• Flows — Flows can reference knowledge
• Guides — Generate guides from documentation