Reply Suggestions
Get AI-generated response options tailored to each ticket. Choose, customize, and send faster with intelligent suggestions.
What Are Reply Suggestions?
Reply suggestions are AI-generated response drafts based on:
- The ticket content and context
- Your knowledge base articles
- Your organization's tone and style
- Similar past tickets
Example
Customer message: "I ordered a scooter 2 weeks ago and still haven't received it. Order #12345."
AI Suggestions:
-
Apologetic + Action
"I'm sorry for the delay with your order #12345. Let me check the shipping status right away and get back to you within the hour with an update."
-
Information Request
"Thank you for reaching out about order #12345. To help track this down quickly, could you confirm the shipping address on your order?"
-
Resolution Focused
"I apologize for the wait on order #12345. I've just checked and your package is currently at [tracking status]. It should arrive by [date]. I'll personally monitor this to ensure it reaches you."
Accessing Suggestions
From the Sidebar
- Open a ticket
- Look for "Reply Suggestions" in the AI sidebar
- Click to expand and view options
Keyboard Shortcut
Press Cmd/Ctrl + R to generate/show reply suggestions.
From Composer
- Click the ✨ icon in the reply composer
- Suggestions appear as options
- Click to insert
Using Suggestions
Insert a Suggestion
- Review the available options
- Click the suggestion you prefer
- Text is inserted into the composer
- Edit as needed
- Send
Copy to Clipboard
- Hover over a suggestion
- Click the copy icon
- Paste wherever needed
Regenerate Options
If suggestions aren't quite right:
- Click "Regenerate" or refresh icon
- New suggestions generated
- Based on same context
Suggestion Types
Tone Variations
Suggestions often vary by tone:
| Type | When to Use |
|---|---|
| Professional | B2B, formal situations |
| Friendly | Consumer support, casual brands |
| Apologetic | When customer is frustrated |
| Direct | Quick answers, simple questions |
Content Variations
Suggestions may focus on different approaches:
| Approach | Description |
|---|---|
| Action-oriented | "I'll fix this right now" |
| Information-gathering | "Can you tell me more about..." |
| Educational | "Here's how this works..." |
| Escalation | "I'm bringing in our specialist..." |
How Suggestions Are Generated
Context Sources
AI considers:
- Current ticket: Subject, description, messages
- Customer history: Past interactions
- Knowledge base: Relevant articles
- Similar tickets: How you've responded before
- Organization style: Your tone preferences
Generation Process
Ticket Content → AI Analysis → Draft Generation → 3 Options Presented
Customizing Your Suggestions
Organization Settings
In Settings → AI → Reply Suggestions:
- Tone preference: Professional, Friendly, Balanced
- Signature inclusion: Auto-add signatures
- KB priority: Prefer KB content in suggestions
- Response length: Concise vs. detailed
Per-Ticket Customization
Click "Customize" when generating to specify:
- Specific tone for this ticket
- Topics to include/exclude
- Length preference
Best Practices
Use as Starting Points
Suggestions are starting points, not final answers:
- Review the suggestion
- Personalize for the customer
- Add specific details
- Verify accuracy
- Then send
Maintain Authenticity
- Add your personal touch
- Don't sound robotic
- Customers appreciate human connection
Verify Before Sending
Always check:
- Technical accuracy
- Current policies
- Customer-specific details
- Tone appropriateness
When Suggestions Excel
Great For
- Common questions: FAQs, standard requests
- Initial responses: Getting started quickly
- Writer's block: When you're stuck
- Consistency: Maintaining brand voice
- Training: Showing new agents good responses
Less Ideal For
- Complex technical issues: Need expert input
- Highly emotional situations: Need personal touch
- Legal/compliance matters: Need careful wording
- Unique edge cases: Need custom solutions
Suggestion Quality
Improving Quality
Better suggestions come from:
- Rich knowledge base: More articles = better matches
- Consistent responses: AI learns from patterns
- Clear ticket content: Good input = good output
- Customer context: History informs suggestions
Quality Indicators
Suggestions show confidence levels:
- ⭐⭐⭐ High confidence
- ⭐⭐ Medium confidence
- ⭐ Lower confidence (may need more editing)
Usage & Limits
Plan Limits
| Plan | Monthly Suggestions |
|---|---|
| Free | 50 |
| Pro | 500 |
| Business | Unlimited |
Checking Usage
View usage in Settings → Usage → AI Features.
Troubleshooting
"No suggestions available"
- Ticket may be too short
- Try adding context and regenerating
- Check AI settings are enabled
"Suggestions seem generic"
- Build out your knowledge base
- Add more context to tickets
- AI improves over time
"Suggestions don't match our tone"
- Configure tone in Settings → AI
- Provide examples of good responses
- Regenerate after settings change
"Same suggestions every time"
- Try the "Regenerate" button
- Add more ticket context
- Different tickets should yield different suggestions
Tips for Maximum Value
- Use suggestions for speed: Don't write from scratch
- Always personalize: Add customer name, specific details
- Learn from suggestions: See patterns in good responses
- Provide feedback: Good suggestions improve the system
- Build your KB: More knowledge = better suggestions