AI Server Actions: Intelligent Automation for Odoo
Automatically update record fields using AI. Generate descriptions, summaries, and enrich content without writing code.
Key Features
Key points
• AI-Powered Field Updates: Select "Update with AI" instead of fixed values to let AI automatically fill content based on your prompt.
• Built-in Server Actions Integration: Works within Odoo's "Update Record" action type, no additional modules needed.
• Multiple Field Type Support: Compatible with Char, Text, Html, and other stored field types.
• Natural Language Prompts: Write simple requests like "Generate a professional summary" instead of complex code.
• Full Automation: Combine with Automation Rules to update records automatically when events occur.
Important Terminology
Key points
• Server Action: Automated action executed on Odoo server when trigger conditions are met.
• Value Type: Method of assigning values to fields - can be fixed value, Python code, or "Update with AI".
• Update with AI: Special mode allowing AI to automatically generate field content based on prompts.
• AI Update Prompt: Natural language instruction describing what content AI should generate.
• Stored Field: Field with data saved directly to database, required for AI updates.
• Action To Do: Type of action the server action will perform - must be "Update Record" to use AI.
Step-by-Step Usage
Key points
• Step 1 - Access Server Actions: Go to Settings → Technical → Automation → Server Actions. Enable Developer Mode if Technical menu is not visible.
• Step 2 - Create or Edit Action: Create new or open existing server action. Select the target model (e.g., res.partner, sale.order).
• Step 3 - Configure Update Record: Set "Action To Do" to "Update Record". Choose target field in "Fields to Update" section.
• Step 4 - Select Update with AI: At "Value Type", choose "Update with AI". Interface will display "AI Update Prompt" input field.
• Step 5 - Write Prompt: Enter clear instructions like "Write a professional summary for this customer" or "Generate a concise note describing the issue".
• Step 6 - Save and Test: Save configuration. Test on sample records to verify results before applying to production automation.
Real-World Use Cases
Key points
• Auto-Generate Product Descriptions: When creating new products, AI automatically generates marketing descriptions based on name and attributes.
• Summarize Customer Requests: When receiving contact forms or emails, AI creates concise summaries for staff to quickly understand issues.
• Enrich Partner Information: Automatically generate internal notes about customers based on transaction history and interactions.
• Classify and Tag Content: AI reads ticket/lead content and automatically fills appropriate category or priority fields.
• Generate Email Content: Create professional email content based on templates and record information.
• Standardize Data: Automatically format and standardize text fields according to company standards.
Effective Prompt Writing Tips
Key points
• Be Specific: Instead of "Write something", use "Write a short professional summary of the customer request in 2-3 sentences".
• Define Style: Add requirements like "in a professional tone", "keep it concise", "use friendly language".
• Set Clear Limits: "Use maximum 100 words", "Return a single sentence", "Keep under 50 characters".
• Start with Simple Text Fields: Experiment with Char, Text fields before applying to more complex fields.
• Test on Real Data: Run action on sample records, review results, adjust prompt accordingly.
• Separate Multiple Goals: If updating multiple fields with different logic, create separate server actions.
Conditions and Limitations
Key points
• Stored Fields Only: Field must be stored in database. Non-stored computed fields will be blocked with "not stored" error.
• Requires Update Record Type: Only works when "Action To Do" is "Update Record", not applicable to other action types.
• Write Permissions Required: Action must run in context with permissions to update target records.
• Depends on ai_fields Module: ai_server_actions module requires ai_fields module to be installed and active.
• Results Depend on Prompt Quality: Unclear prompts will produce incorrect or overly generic results.
Common Troubleshooting
Key points
• "Update with AI" Not Visible: Check that ai_server_actions and ai_fields modules are installed, ensure you're in Update Record mode.
• Error "field can not be computed with AI": Selected field is not a stored field. Choose different field or modify field definition.
• AI Results Not as Expected: Write more specific prompts, add length/style requirements, retest on sample records.
• Action Runs But No Update: Check trigger conditions, verify correct target record, check write permissions.
• Content Too Long or Short: Add explicit limits to prompt like "maximum 100 words" or "at least 3 sentences".
Best Practices
Key points
• Start with Description Fields: Use AI for notes, descriptions, summaries before applying to critical data.
• Avoid Sensitive Data: Don't use AI directly for financial or legal data without additional validation layers.
• Keep Prompts Short and Clear: Good prompts are typically short, have specific goals, and clear output constraints.
• Test Thoroughly Before Deployment: Run tests in staging environment with sample data before production use.
• Monitor and Optimize: Track AI results, collect feedback, adjust prompts to improve accuracy.
• Prepare Prompt Templates: Create library of standard prompts for each business process type for reuse.