AI in support workflows¶
Utilizing Odoo’s AI features in the Helpdesk app assists support agents with their tasks, rather than replacing existing workflows. The goal is to reduce repetitive tasks, accelerate ticket handling, and improve consistency by allowing AI to analyze ticket content and suggest or execute actions automatically.
Instead of manually categorizing, prioritizing, and responding to every request, support teams can rely on AI to interpret incoming information and perform preliminary processing. Agents remain in control of decisions and customer communication, while AI handles repetitive analysis and drafting.
How AI works in Helpdesk¶
AI capabilities in Helpdesk are built around three core ideas:
AI agents that interact with users or agents and generate responses based on knowledge sources.
AI automations that update tickets automatically when records are created or modified.
AI enabled fields that generate structured information, such as summaries, from unstructured text.
AI agents¶
An AI agent is a configurable assistant that operates using prompts and contextual knowledge. The agent itself does not contain business logic; instead, it follows instructions defined by users and relies on provided documentation.
When an agent receives a request, it combines its system prompt, the current record context (such as a ticket), and its knowledge sources.
This allows the agent to provide responses grounded in internal documentation rather than generic language model output.
Agents can be restricted to specific sources or granted permissions to perform actions using topics. This ensures that support assistants behave predictably and remain aligned with organizational rules.
AI automations¶
Automation rules allow AI to act during record lifecycle events. Instead of waiting for an agent to categorize or prioritize tickets, AI can perform these tasks automatically as soon as a ticket is created.
The key principle is that AI does not understand Odoo models directly. It receives text extracted from fields and produces an output based on prompt instructions. Users guide AI behavior through clear prompting and carefully selected input fields.
AI-velden¶
AI enabled fields extend standard Odoo fields by allowing their values to be generated automatically through prompts.
These fields are useful when large amounts of text need to be condensed into actionable information. For example, a long customer message can automatically be summarized into a short list of key points, allowing agents to understand issues quickly without reading full descriptions.
Configuratie¶
Configuring an AI agent for support¶
AI agents are configured in the AI application. To configure an existing agent, navigate to and click the (vertical ellipsis) icon, then click Configuration on the drop-down menu.
Configuring begins by selecting a language model and defining a system prompt describing how the agent should behave. The prompt determines tone, scope, and responsibility. A support-focused agent might be instructed to analyze issues and provide concise technical responses.
The next step is defining the agent’s sources. These sources form the agent’s memory and should contain trusted documentation such as product guides or FAQs. The quality of these sources directly impacts the quality of AI responses.
Tip
Select Restrict to Sources to require the agent to only base their responses on the supplied resources.
Users may also define topics, which grant the agent permission to perform specific actions. Topics allow advanced scenarios such as record creation or automation triggers.
Once an agent exists, default prompts can be used to initiate a conversation.
A default prompt defines where and when the AI appears inside Odoo. In the Helpdesk app, this typically means allowing the AI button to appear on ticket records. The default prompt specifies the model where it is available, which agent is called, and additional contextual instructions.
These instructions act as a second layer on top of the agent prompt, adapting the agent’s behavior to Helpdesk scenarios.
Configuring AI automations¶
AI server actions are configured using standard automated actions, with AI used as the update mechanism.
When defining a server action, users specify when the action should run, such as when a new ticket is created. The AI prompt then explains how the ticket should be interpreted and which field should be updated.
This approach keeps automation predictable while allowing AI to make contextual decisions.
Configuring AI fields¶
AI fields are configured by enabling the AI option on a custom or property field and providing a prompt describing the expected result. Once configured, the field updates automatically when the record changes, giving agents instant access to structured information.
Waarschuwing
AI fields can be added to a record through the Studio app or property field. Installing Studio may impact the current pricing plan for a database.
For more information, refer to Odoo’s pricing page or contact your account manager.
Zie ook