AI Assistants Guide¶

AI assistants are virtual agents that users interact with in your Odoo system. They combine AI provider configurations, behavioral contexts, and knowledge bases to provide intelligent responses to user queries.

🤖 Overview¶

AI assistants are the core interface between users and AI capabilities. Each assistant is configured with:

AI Provider: The underlying AI model (ChatGPT, Claude, Gemini, OdooAI)
Behavior Context: Personality and response style
Knowledge Base: Data sources for accurate information
Access Control: Permissions and user group restrictions

🧠 Understanding AI Assistants¶

How AI Assistants Work¶

User Interaction: Users chat with assistants through Discuss or Livechat
Query Processing: Assistant analyzes the user's question
Knowledge Retrieval: Searches data sources for relevant information
Context Integration: Combines user query with retrieved knowledge
AI Response: Generates response using the configured AI provider
Response Delivery: Sends response back to the user

Assistant Capabilities¶

Core Functions

Answer questions based on configured knowledge
Provide information from data sources
Maintain conversation context
Handle multiple conversation threads
Access Odoo records when permitted
Create, update, and delete records with proper validation

Advanced Features

File attachment processing (varies by provider)
Record reference understanding ($model/id)
Semantic search across knowledge base
Data modification operations (create, update, delete)
Method discovery and execution for custom business logic
Read conversations and post messages to Odoo threads
Context-aware responses
Human handover capabilities (Livechat)

🛠️ Creating AI Assistants¶

Basic Setup¶

Access AI Assistants
- Go to Settings → Technical → AI → Assistants
- Click Create to add a new assistant
Basic Information
- Name: Assistant's name (e.g., "Customer Support Bot")
- Description: Brief description of capabilities
- Sequence: Display order (lower numbers appear first)

Configuration Settings¶

AI Provider Configuration

Configuration: Select the AI provider configuration
Type: Automatically set based on the configuration
Model: AI model to use (auto-filled from configuration)

Behavior Context

Context: Choose the behavioral context
Defines personality and response style
Controls how the assistant behaves

Access Control¶

Super Assistant Mode

Is Super Assistant: Enable for full system access
When enabled, assistant can access all system models and data sources
Useful for administrative assistants
Use with caution for security

Group Restrictions

Groups: Limit access to specific user groups
Leave empty for public access
Useful for department-specific assistants
Controls who can see and use the assistant

Data Source Assignment

Data Sources: Assign specific knowledge bases
Assistant can only access assigned data sources
Multiple data sources can be assigned
Choose relevant sources for the assistant's purpose

Note: For data sources of Model type, AI can only operate within user's permission limit, even in Super Assistant mode.

Data Modification Capabilities¶

AI assistants can perform data modification operations on Odoo records:

Create Operations:

Create new records in accessible models
Automatic domain validation after creation
Only creates records within allowed domain

Update Operations:

Update existing record fields
Domain access validation before update
Only updates records within allowed domain

Delete Operations:

Delete records with proper validation
Domain access validation before deletion
Only deletes records within allowed domain

Security:

All operations respect user permissions
Domain restrictions enforced
Field-level access control

Advanced Method Capabilities¶

AI assistants can discover and execute custom business logic methods on Odoo records. This allows AI to perform complex operations beyond simple create, update, and delete.

Method Discovery:

AI can discover what methods are available on any accessible model
Shows method names, parameters, and documentation
Helps AI understand what operations are possible on records

Method Execution:

AI can execute custom business logic methods on specific records
Supports both simple and complex method calls with parameters
Only methods that work on records (not model-level methods)
Automatic validation ensures security and data integrity

Security:

All method calls respect user permissions
Domain restrictions checked before and after execution
Field-level access control ensures only allowed fields can be changed
Blacklisted methods cannot be called (use dedicated tools instead)

Thread Communication:

Conversation Reading: AI can read conversation history of any accessible Odoo thread
Message Posting: AI can send messages to any Odoo thread (tasks, leads, orders, etc.)
Support for attachments, notifications, and different message types
Automatic permission checking ensures secure communication

Field Translation:

Get Translations: AI can retrieve current translations of translatable fields
Update Translations: AI can update translations for multiple languages
Support for both simple translation and term-based translation
Automatic handling of translation formats based on field type

🎯 Assistant Types and Use Cases¶

Customer Support Assistant¶

Purpose: Handle customer inquiries and support requests

Configuration:

Name: "Customer Support Bot"
Context: Customer service focused
Data Sources: Product catalog, FAQ, support documentation
Access Groups: Public or customer groups

Capabilities:

Answer product questions
Provide support information
Handle basic troubleshooting
Escalate to human agents

Internal Assistant¶

Purpose: Help internal teams with business processes

Configuration:

Name: "Internal Assistant"
Context: Professional, internal knowledge
Data Sources: Company policies, procedures, internal docs
Access Groups: Employee groups only

Capabilities:

Answer process questions
Provide policy information
Help with data analysis
Assist with documentation

Specialized Assistant¶

Purpose: Domain-specific expertise

Configuration:

Name: "Technical Support", "Sales Assistant", etc.
Context: Technical or specialized behavior
Data Sources: Domain-specific knowledge
Access Groups: Relevant department groups

Capabilities:

Technical troubleshooting
Sales information
Domain-specific guidance
Specialized knowledge access

Administrative Assistant¶

Purpose: System-wide access for administrators

Configuration:

Name: "Admin Assistant"
Context: Administrative, system-focused
Is Super Assistant: Enable for broad system access without manual configuration
Access Groups: Administrator groups

Capabilities:

Access to all system models and data sources
Administrative tasks and system monitoring
Advanced troubleshooting and data analysis
Cross-company operations when user has permissions
Still respects user's individual data access rights

⚙️ Advanced Configuration¶

Assistant Behavior¶

Context Selection

Choose context that matches assistant's purpose
Consider tone and style requirements
Test different contexts for best results
Customize context for specific needs

Response Quality

Monitor assistant responses
Adjust context if needed
Test with various scenarios
Ensure consistent quality

Knowledge Management¶

Data Source Selection

Choose relevant data sources
Consider assistant's purpose
Balance coverage vs. performance
Regular review and updates

Access Control

Limit access appropriately
Use group restrictions
Monitor usage patterns
Regular security reviews

🔗 Integration Setup¶

Discuss Integration¶

Automatic Setup

Assistants automatically appear in Discuss sidebar
Users can start private chats
Group conversations can mention assistants
Record integration available

User Experience

Easy access through sidebar
Intuitive chat interface
Context preservation

Livechat Integration¶

Channel Assignment

Assign assistants to livechat channels
Mix AI and human operators
Configure escalation logic

Operator Management

AI handles initial inquiries
Human operators for complex cases
Automatic handover capabilities

📋 Best Practices¶

Assistant Design¶

Clear Purpose

Define specific use case
Choose appropriate context
Select relevant data sources
Test thoroughly

User Experience

Use clear, descriptive names
Provide helpful descriptions
Ensure consistent behavior
Monitor user satisfaction

Performance Optimization¶

Data Source Management

Choose relevant sources only
Monitor response quality
Update sources regularly
Optimize for speed

Access Control

Limit access appropriately
Use group restrictions
Monitor usage patterns
Regular security audits

Quality Assurance¶

Testing

Test with various scenarios
Verify response accuracy
Check context appropriateness
Monitor user feedback

Monitoring

Track usage patterns
Monitor response quality
Review error logs
Regular performance reviews

🔧 Troubleshooting¶

Common Issues¶

Assistant Not Responding

Check configuration is active
Verify AI provider is working
Ensure context is properly set
Check data source accessibility

Poor Response Quality

Review and improve context
Check data source quality
Verify AI provider settings
Test with different scenarios

Access Issues

Check group restrictions
Verify data source permissions
Review super assistant settings
Test with different users

Performance Issues¶

Slow Responses

Check AI provider performance
Review data source size
Optimize context length
Monitor system resources

Memory Issues

Reduce data source size
Optimize field selections
Monitor database usage
Regular maintenance

🔄 Maintenance¶

Regular Tasks¶

Content Updates

Update data sources regularly
Review and improve contexts
Monitor assistant performance
Update configurations as needed

Security Reviews

Review access permissions
Monitor usage patterns
Update group assignments
Regular security audits

Performance Monitoring

Track response times
Monitor error rates
Review user feedback
Optimize configurations

This guide covers the essential aspects of AI assistant creation and management. Well-configured assistants provide significant value to users while maintaining security and performance.