Kanbox integrates AI directly into your outreach campaigns to automate, qualify, and manage conversations at scale. This system allows you to combine automation with intelligent decision-making, ensuring relevant and personalized interactions.
This documentation explains:
How to configure the AI agent
How AI tracks and manages conversations in the pipeline
1. Configuring the AI Agent
The AI agent is configured at the campaign (sequence) level and defines how conversations are handled automatically.
1.1 Enable AI Conversations
In your sequence settings:
Activate “AI Response”
Choose the AI mode:
Assist → AI suggests replies, user validates
Autopilot → AI replies automatically
1.2 Define the Conversation Goal
This is the most critical part of the setup.
The Conversation Goal tells the AI what outcome to drive toward.
Example:
Book a meeting
Qualify a lead
Send a resource
Best practices:
Be explicit and action-oriented
Include links if needed (calendar, landing page, etc.)
Add constraints to guide behavior
Example configuration:
Book a meeting using a calendar link
Only push for booking if the lead shows interest
If not interested, do not insist
👉 This ensures the AI behaves like a human SDR, not a spam bot.
1.3 Provide Company / Offer Context
This section gives the AI the necessary background to answer intelligently.
Include:
What your company does
Value proposition
Optional: objections handling elements
Why it matters:
Improves response relevance
Ensures consistency in messaging
Reduces hallucinations
👉 Think of this as training the AI on your product.
1.4 AI Behavior Logic
Based on your configuration, the AI will:
Understand user intent (interest, objection, question, etc.)
Adapt tone and responses dynamically
Push toward the defined goal when relevant
Stop or redirect if the user is not a good fit
It can:
Answer questions
Qualify leads
Handle objections
Propose meetings
2. AI Tracking in the Pipeline
Kanbox provides full visibility into AI-driven conversations through a Kanban pipeline.
Each conversation evolves through stages based on user behavior and AI analysis.
2.1 Pipeline Structure
2 new columns include:
AI Conversation → ongoing conversations handled by AI
To review → user replied, waiting or processed
2.2 AI Insights (Intent Detection)
Each conversation is enriched with AI Insights, which classify user intent in real time.
Examples of detected intents:
Interested
Meeting intent
Objection
Later
These labels appear directly on conversation cards.
2.3 Conversation Status Tracking
Each lead card includes:
AI Insight label (intent classification)
Progress indicator (e.g. “0/3”, “1/3”)
Status (e.g. Pending)
This allows you to:
Understand where each lead is in the funnel
Prioritize high-value conversations
Identify bottlenecks
2.4 Human + AI Collaboration
Even in Autopilot mode, you maintain control:
Monitor all conversations
Jump in at any time
Adjust strategy based on insights
Kanbox ensures:
Automation handles volume
Humans handle edge cases and strategy
Key Benefits
Scalable outreach with personalized conversations
Automated qualification of leads
Real-time intent detection
Clear pipeline visibility
Higher conversion rates with less manual work