Persistence and Rebuttal Settings
The key difference between Site2CRM AI and basic chatbots: our AI doesn't give up easily.
Why Persistence Matters
Most website visitors say "no thanks" reflexively. A basic chatbot accepts this immediately. But studies show:
- 80% of sales require 5+ follow-ups
- Most reps give up after 1-2 attempts
- Persistent (but respectful) follow-up converts more leads
Persistence Levels
| Level | Description | Best For |
|---|---|---|
| Soft | 1-3 gentle attempts, backs off quickly | Support, low-pressure brands |
| Medium | Up to 5 attempts, balanced approach | Most businesses (recommended) |
| Aggressive | Full attempts, multiple angles | High-value sales, limited-time offers |
Setting Persistence Level
- Go to Chat Widget > select your agent
- Find Behavior Settings section
- Choose Persistence Level
- Adjust Rebuttal Count (1-10)
- Save changes
Rebuttal Count
This is the maximum number of times the AI will try to re-engage after an objection:
- 1-2: Very light touch, accepts "no" quickly
- 3-5: Balanced, tries different angles (recommended)
- 6-10: Very persistent, for high-value conversions
How Rebuttals Work
When a visitor says "no thanks," "not interested," or similar:
Attempt 1: Add value "Totally understand! Quick thing though - [benefit]. Would that be helpful?"
Attempt 2: Reduce risk "I hear you! It's completely free to try. What email should I send the link to?"
Attempt 3: Understand objection "Fair enough! What's specifically holding you back? I might be able to help."
Attempt 4: Offer alternative "No problem! Would a quick 2-minute call work better?"
Attempt 5: Final soft push "Totally respect that! How about a one-page overview for later? What email?"
After max attempts: Accept gracefully "No worries at all! We're here whenever you're ready."
Tips
Don't set rebuttal count too high
- More than 7-8 can feel annoying
- Quality over quantity
Match persistence to your brand
- Luxury brands: softer approach
- Discount/deal brands: more aggressive OK
- Support contexts: always soft
Test and iterate
- Start with medium/5 rebuttals
- Check conversation logs
- Adjust based on feedback