Model
Last updated on March 30, 2026
The AI Model Settings page allows you to fine-tune the intelligence, strictness, and verbosity of your AI agent. These controls let you balance response speed, accuracy, and detail to fit your specific support needs.
Key Features
- AI Model Selection: A dropdown to select the underlying AI model, allowing Premium users to upgrade to a more advanced model.
- Match Threshold: A sliding scale to control how strictly the AI matches user queries to your knowledge base documents.
- Reasoning Effort (GPT-5): A dropdown to adjust the cognitive effort the model spends before answering complex queries.
- Verbosity: A dropdown to instruct the AI on how long and detailed its responses should be.
Selecting the AI Model (Premium Feature)
- Locate the AI Model dropdown.
- GPT-5 Nano (Standard): The default model used for standard support interactions.
- GPT-5.4 Nano (Advanced): Offers faster speeds and better reasoning capabilities. This option requires a Premium plan.
Adjusting the Match Threshold
- Locate the Match Threshold slider.
- Move the slider to a lower value to make the bot "broader," pulling more documents into its context window.
- Move the slider to a higher value to make the bot "stricter," pulling fewer documents and only answering when it is highly confident.
Setting Reasoning Effort (GPT-5)
- Use the Reasoning Effort (GPT-5) dropdown to determine how much time the AI spends "thinking" before generating a reply:
- Low (Faster): Best for simple, straightforward questions where speed is the priority.
- Medium (Balanced): The default setting, providing a good mix of speed and comprehension.
- High (Deep thinking): Best for complex, multi-step user queries that require careful logical deduction.
Controlling Verbosity
- Use the Verbosity dropdown to set the default length of the AI's answers:
- Low (Concise): Forces the AI to give brief, straight-to-the-point answers.
- Medium: The standard conversational length.
- High (Detailed): Encourages the AI to provide comprehensive, thorough, and highly detailed explanations.
Important Notes & Pro Tips
- Permissions: Only workspace owners can modify the AI Model settings. If you are logged in as an agent, these controls will be disabled (grayed out).
- Premium Model Access: The GPT-5.4 Nano (Advanced) model is exclusively available to workspaces on the Premium plan. For free users, the selector is disabled, locked to the standard model, and displays a "Premium" badge.
- Premium Fair Use Limit: If you are using the Premium GPT-5.4 Nano model, there is a zero-latency fair use limit of 10,000 monthly messages. If your workspace exceeds this limit within a given month, the system will automatically fall back to the standard GPT-5 Nano model to ensure uninterrupted chat service.
Related Articles
Behavior
The Behavior page allows you to customize how your AI agent interacts with customers. From defining its persona to setti...
BehaviorSettings
Knowledge Base
The Knowledge Base page is where you train your AI agent. By adding data sources like websites, files, and third-party t...
Knowledge BaseData Sources
Was this helpful?
