Fine-tuning
Fine-tuning
Fine-tuning rules in FirstQuadrant allow you to customize how the AI behaves across various sales scenarios. These rules serve as high-level instructions—like guidelines for a smart assistant—not strict logic trees. They enable FirstQuadrant to reason through nuanced sales interactions while adapting to your unique sales process and language.
What are fine-tuning rules?
Fine-tuning rules are written in plain English and used to:- Teach the AI how to respond to specific messages or notes
- Control how the AI interprets information
- Define sales playbook behavior per contact segment
- Embed sales knowledge and objection handling
Rule types
Fine-tuning rules apply to different content types:| Type | Description |
|---|---|
| Triggers based on labeled inbound emails, e.g. “Out-of-office”, “Counter-offer” | |
| Note | Triggers when a note is added, either manually or via API |
| General rule | Provides overarching behavioral guidance |
| Knowledge | Stores reusable sales knowledge for recurring questions |
Best practices
- Start with the default rules
FirstQuadrant comes with pre-set rules that work well for most users. Customize them only once you’re familiar with how the AI behaves. - Think assistant, not computer
You’re not setting up workflows—you’re training a reasoning assistant. - Segment by contact types
You can assign different rules for different customer segments like “buyers” vs “sellers.” - Iterate gradually
Add or tweak rules only when the AI’s behavior doesn’t match your expectations.
Managing rules
All existing rules are displayed in a centralized list under the Fine-tuning section. From here, you can:- View all labels and types
- See at-a-glance instructions
- Edit or update rules as needed
Previous
Create fine-tuningFine-tuning rules in FirstQuadrant allow you to customize the AI's behavior by defining how it should respond in specific scenarios. Rather than hard-coded if-statements, these rules act as flexible, natural-language instructions—like training an assistant. You can create fine-tuning rules for different input types, limit their scope to specific segments, and optionally enable Autopilot to automate actions without human approval.
Next