Knowledge is powerful configuration tools that control how AI behaves and responds according to your specific guidelines. They ensure consistent output and adherence to your standards across all interactions. Knowledge Management Interface

Overview

What are Knowledge?

Knowledge define specific rules and guidelines that AI must follow when processing requests. They can include coding standards, writing styles, security requirements, or any custom behavior rules.

Scoped Control

Knowledge can be scoped to individual users, teams, or entire organizations, providing flexible control over AI behavior at different levels.

Knowledge Structure

Each knowledge contains the following components:
  • Name: A descriptive name for your knowledge
  • Description: Optional description explaining the knowledge’s purpose
  • Rules: The actual guidelines and requirements in text format
  • Organization Level: Applied to all users within an organization
  • Team Level: Applied to all members of a specific team
  • Individual Level: Personal knowledge for individual users

Creating Knowledge

1

Define Your Rules

Write clear, specific guidelines that you want the AI to follow. These can include:
  • Response formatting standards
  • Writing tone and style requirements
  • Internal Tribal Knowledge, Policies, Acronyms, etc.
  • Domain-specific requirements
2

Set the Scope

Choose whether the knowledge applies to:
  • Individual: Personal use only
  • Team: All team members (main)
  • Organization: Everyone in your organization
3

Activate the Knowledge

Once created, knowledge can be activated and applied to AI interactions automatically based on their scope. You can select knowledge using the dropdown in the main chat page.

Best Practices

Keep rules specific and actionable. Vague rules like “be creative” are less effective than specific guidelines like “use TypeScript interfaces for all data structures.”
Avoid conflicting rules. When multiple knowledge are active, ensure they don’t contain contradictory requirements.

Writing Effective Rules

  1. Be Specific: Use concrete examples and clear requirements
  2. Use Positive Language: Focus on what to do rather than what not to do
  3. Include Context: Explain why certain rules exist
  4. Test Thoroughly: Verify that rules produce the desired behavior

Knowledge Examples

Development Standards
# Code Quality Requirements

- Use TypeScript for all new code
- Follow ESLint configuration strictly  
- Include JSDoc comments for all functions
- Use kebab-case for file names
- Implement proper error handling with try/catch
- Write unit tests for all new functions
- Use meaningful variable and function names
- Limit function length to 50 lines maximum
Content Standards
# Documentation Style Guide

- Use active voice in all documentation
- Keep sentences under 20 words when possible
- Include code examples for all technical concepts
- Use consistent terminology throughout documents
- Follow AP style guide for formatting and grammar
- Start with action verbs in step-by-step instructions
- Include relevant cross-references and links
- End sections with clear next steps
Security Standards
# Security Best Practices

- Never expose API keys or secrets in code
- Use environment variables for all configuration
- Implement input validation for all user data
- Follow OWASP Top 10 security guidelines
- Require authentication for all sensitive operations
- Use HTTPS for all external communications
- Sanitize all user inputs before processing
- Log security events for audit purposes
Brand Standards
# Brand Voice and Tone

- Use friendly, professional tone in all communications
- Avoid technical jargon when speaking to non-technical users
- Include Context branding in appropriate contexts
- Use consistent color schemes (#6366F1 primary, #818CF8 light)
- Follow UI/UX design system for all interface elements
- Maintain consistent spacing and typography
- Use approved imagery and visual assets only

Managing Knowledge

Activation and Priority

  • Knowledge is automatically applied based on their scope
  • Organization knowledge have the highest priority
  • Team knowledge override individual preferences
  • Multiple knowledge can be active simultaneously

File Attachments

Knowledge support file attachments for additional context:
  • Style guides (PDF, Markdown)
  • Code examples (any programming language)
  • Templates (document templates, boilerplates)
  • Reference materials (specifications, documentation)
Attached files are parsed and their content is included in the AI’s context when the knowledge is active.

Integration with Workflows

Knowledge work seamlessly with workflows to provide comprehensive AI guidance:
  • Active Knowledge: Always applied during workflow execution
  • Workflow-Specific Rules: Some workflows may have embedded rule requirements
  • Scope Inheritance: Workflows inherit applicable knowledge based on their scope