Intelligent Response Mode

Intelligent Response Mode (IRM) is a smart email response feature in Eloquens AI that automatically decides whether to send your AI-generated email responses immediately or create drafts for your review. Think of it as having an intelligent assistant that knows when it’s confident enough to respond on your behalf and when it should ask for your approval first.

How to Set Up IRM

1. Choose Response Mode

When creating or editing a campaign, you’ll see three response mode options:

• Draft: AI always creates drafts for your review
• Send: AI automatically sends all responses
• Intelligent Response: AI decides between sending and drafting based on smart rules

Select “Intelligent Response” to enable IRM.

2. Configure Smart Rules

When you select Intelligent Response, you’ll see a configuration panel titled “When AI decides to create a draft instead of sending:”

The system comes with these built-in smart rules:

• AI Lacking Knowledge (Default: Enabled)
  • Automatically creates a draft when the AI doesn’t have enough information to provide a confident response
  • This is the most important rule and is enabled by default
• Sentiment-Based Rules (Default: Enabled)
  • Creates drafts for messages with specific emotional tones
  • Default examples: “escalations, request to speak to management, extreme dissatisfaction”
  • You can customize this list with your own examples
• Sensitive Topics (Default: Disabled)
  • Creates drafts when messages contain sensitive subjects
  • Examples: “legal, hr, PII”
  • Enable and customize based on your industry needs
• VIP Recipients (Default: Disabled)
  • Creates drafts for messages from important contacts
  • Examples: “ceo@company.com, board@company.com“
  • Perfect for ensuring executive communications get personal attention
• Custom Rule (Default: Disabled)
  • Write your own custom rule in plain language
  • Most flexible option for specific business needs

How IRM Works in Practice

When IRM Chooses to SEND:

• AI has sufficient knowledge to respond confidently
• Message doesn’t trigger any of your enabled rules
• Response goes out automatically without delay

When IRM Chooses to DRAFT:

• AI lacks knowledge about the topic
• Message matches one of your configured rules (sentiment, sensitive topics, VIPs, custom)
• Draft is created and you’re notified for review

Best Practices

✅ Start Conservative

Begin with the default rules enabled:

• AI lacking knowledge ✓
• Sentiment-based rules ✓
• Other rules disabled initially

✅ Customize Gradually

• Monitor which drafts are created
• Add specific terms to sentiment rules based on your experience
• Enable additional rules as needed

✅ Regular Review

• Check your drafts regularly to understand IRM’s decision-making
• Adjust rules based on false positives/negatives
• Fine-tune sensitivity over time

Example Scenarios

Scenario 1: Knowledge Gap

• Customer asks about a new product not in your knowledge base
• IRM → Creates draft (AI lacking knowledge rule triggered)

Scenario 2: Escalation

• Customer says “I want to speak to your manager immediately!”
• IRM → Creates draft (Sentiment rule triggered: “request to speak to management”)

Scenario 3: Routine Inquiry

• Customer asks about business hours (info available in knowledge base)
• IRM → Sends automatically (No rules triggered, AI confident)

Scenario 4: VIP Customer (if enabled)

• Email from ceo@importantclient.com
• IRM → Creates draft (VIP rule triggered)

Monitoring and Adjustment

• Review drafts created by IRM to understand its decision-making
• Look for patterns in when drafts are created vs. sent
• Adjust rules if you notice:
  • Too many drafts (rules too strict)
  • Inappropriate automatic sends (rules too lenient)