Configure SupportBuddy in the flow editor to enable AI-powered suggestions for your support team.
SupportBuddy is configured per flow. Each flow can have its own SupportBuddy with different knowledge sources and settings.
Knowledge collections are the foundation of good suggestions. SupportBuddy searches your selected collections to ground responses in accurate information.
The Behavior section controls how suggestions are written. This includes tone, context, and policies.
Choose how suggestions should sound. You can select a preset or write custom instructions.
| Preset | Description |
|---|---|
| Professional | Clear, courteous, business-appropriate language |
| Friendly | Warm and conversational while remaining helpful |
| Technical | Precise language with relevant terminology |
| Empathetic | Understanding and supportive, acknowledges concerns |
| Concise | Brief and direct, no unnecessary elaboration |
You can also write custom tone instructions, for example: "Be warm but professional. Use simple language. Address the customer by name when known."
Add background information that helps the AI understand your business:
Define what suggestions should NOT include. This is critical for preventing inappropriate responses:
Important: Policies are enforced at generation time, but agents should still review suggestions before sending. Policies reduce risk but don't eliminate it entirely.
Autonomous mode allows SupportBuddy to automatically send high-confidence responses without agent review.
How it works:
Note: This setting enables the option for agents. Agents can still toggle autonomous mode on/off from the browser extension at any time.
Once SupportBuddy is configured, you'll need to share the connection code with your agents:
The connection code is secure and contains encrypted credentials. If you suspect it has been compromised, you can disable and re-enable SupportBuddy to generate a new code.
To temporarily disable SupportBuddy:
You can re-enable at any time by clicking Enable in the modal.