Inbound Agent
24/7 Intelligent Website Engagement That Never Sleeps
Stop losing leads when your team is offline. Inbound Agent deploys AI-powered chat assistants on your website to engage visitors, answer questions, collect information, and qualify leads around the clock—automatically and at scale.
🚀 Why Inbound Agent?
Never Miss a Lead
Respond instantly to website visitors 24/7, even outside business hours, weekends, and holidays.
Instant Qualification
AI agents ask the right questions to identify and prioritize high-intent prospects automatically.
Knowledge-Powered
Agents answer questions accurately based on your custom knowledge base—product docs, FAQs, pricing, and more.
Seamless Handoff
Qualified leads flow automatically to your sales team or Outreach Agent campaigns for immediate follow-up.
How Inbound Agent Works
The Visitor Journey
- Visitor Arrives: A prospect lands on your website (product page, pricing page, or contact page)
- Chat Appears: A friendly chat window invites them to ask questions
- AI Engages: The Inbound Agent responds instantly with relevant, accurate information
- Lead Qualification: Agent asks qualifying questions to understand their needs and fit
- Next Steps: Based on the conversation, agent collects contact info or schedules a meeting
- Auto-Routing: Qualified leads are automatically added to CRM, email sequences, or Outreach campaigns
Conversion at Scale
Inbound Agents handle unlimited concurrent conversations. Whether you have 10 or 1,000 visitors, every single one gets immediate, personalized attention.
Creating Your First Inbound Agent
Step 1: Agent Profile
Set up your agent's identity and voice.
Agent Name:
- Give your agent a first and last name
- Choose a name that fits your brand (e.g., "Sarah Thompson", "Alex Support", or simply "Support")
- The name appears in chat conversations to create a personal feel
Voice Selection:
- Select an AI voice from the preset library
- Preview voices by clicking the play button
- Voice is used for phone interactions (coming soon)
Knowledge Base:
- A knowledge base is automatically created and linked to your agent
- The agent uses this knowledge base to answer questions
- You can update the knowledge base content anytime
One Agent, One Knowledge Base
Each agent is linked to a dedicated knowledge base. This allows you to create specialized agents for different purposes (support, sales, product) with tailored knowledge.
Step 2: Agent Training
Define how your agent behaves, communicates, and handles conversations.
Agent Role Prompt:
This is where you define your agent's personality, responsibilities, and boundaries. Write a prompt (up to 2,000 characters) that describes:
1. Role Definition:
- What is the agent's primary function? (Support, Sales, Product advisor)
- What department or team does it represent?
2. Responsibilities:
- What questions should the agent answer?
- What information should it provide?
- What data should it collect from visitors?
3. Conversation Style:
- Tone: Friendly, professional, formal, casual
- Language: Simple, technical, industry-specific
- Response length: Concise or detailed
4. Behavioral Boundaries:
- What topics should the agent NOT discuss?
- How to handle questions outside its knowledge?
- When to escalate to human support?
Example Agent Role:
You are a friendly customer support assistant for [Company Name]. Your responsibilities include:
- Answering product feature questions based on our documentation
- Helping users troubleshoot common issues
- Collecting contact information from interested prospects
Your communication style:
- Use a friendly, professional tone
- Keep responses clear and concise (2-3 sentences when possible)
- Base all answers on the knowledge base content
- If unsure, recommend contacting our human support team
You should NOT:
- Promise specific pricing or discounts
- Provide product roadmap information
- Access or request sensitive user data
Auto-Enhance Feature:
- Click "Auto-Enhance" to have AI optimize your prompt
- The system expands and refines your instructions
- Review the enhanced version and modify as needed
- Can be used multiple times to iterate
Writing Effective Prompts
Be specific about what your agent should and shouldn't do. Clear boundaries prevent confusion and maintain consistent quality. Include examples of good responses if helpful.
Step 3: Agent Test and Deploy
Test your agent, customize the chat window, and get ready to deploy.
Chat Window Customization
Chat Window Title:
- Set the title displayed at the top of the chat window
- Default: "Inbound Agent Chat"
- Recommendations: "Live Support", "How can we help?", "Chat with us"
- Keep it inviting and brand-friendly
Chat Window Color:
- Choose your brand's primary color for the chat widget
- Entered as a hex color code (e.g., #0d78b0)
- This color appears on the chat button, header, and send button
- Match your website's color scheme for visual consistency
Test Your Agent
Before deploying to your website, test the agent's responses:
Testing Interface:
- A chat panel appears on the right side of the screen
- Send test messages to see how your agent responds
- Verify response quality, accuracy, and tone
- Check that the agent follows your role instructions
What to Test:
- Common Questions: Ask typical visitor questions (pricing, features, support)
- Edge Cases: Test questions outside the agent's knowledge scope
- Tone and Style: Verify the conversational style matches your brand
- Response Time: Confirm responses are delivered quickly
- Handoff Scenarios: Test how the agent handles escalation requests
Iterative Testing
Test, refine your prompt, test again. Most effective agents are refined through multiple iterations based on real conversation patterns.
Get Your Embed Code
Once you're satisfied with your agent's performance:
- Navigate to the "Widget" section in Step 3
- Click the "Copy" button next to the embed code
- The code is copied to your clipboard with a "Copied!" confirmation
- The code contains your unique Agent ID and API Key pre-filled
Ready to Deploy
Your embed code is a simple snippet that you'll paste into your website. See the "Website Integration" section below for detailed installation instructions.
Phone Number (Coming Soon)
Voice Inbound Capability:
- Purchase a dedicated phone number for your agent
- Visitors can call and speak with your AI agent
- Voice recognition converts speech to text
- Agent responds using the selected voice
- Creates a complete omnichannel support experience
Coming Soon
Phone number functionality is currently in development. When available, you'll be able to add voice capabilities to your Inbound Agent for phone-based interactions.
Managing Your Inbound Agents
Agent Dashboard
Access your agents at: AI SDR → Inbound Agent
Dashboard Overview:
- View all your Inbound Agents in a card layout
- Each card displays agent name, status, voice, and phone number (if purchased)
- Easily see which agents are active or inactive
- Create new agents or manage existing ones
Agent Status
Active / Inactive Toggle:
- Each agent card has an Active/Inactive switch
- Active: Agent is live and responding to website visitors
- Inactive: Agent is paused and won't respond to new conversations
- Toggle instantly without reloading the page
When to Pause:
- Temporarily disable during website maintenance
- Pause while updating knowledge base content
- Disable agents that are no longer needed without deleting them
Agent Actions
Click the three-dot menu on any agent card:
Edit:
- Opens the agent configuration interface
- Modify agent name, role prompt, or voice
- Update chat window title and color
- Changes take effect immediately
Delete:
- Permanently remove an agent
- Displays a confirmation dialog before deletion
- Cannot be undone—deleted agents are permanently removed
- Associated conversations remain in analytics
Copy Embed Code:
- Quick access to copy your agent's website embed code
- Click and the code is copied to clipboard
- "Copied!" confirmation appears
Deletion Warning
Deleting an agent is permanent and cannot be undone. If you only need to temporarily disable an agent, use the Active/Inactive toggle instead.
Website Integration
Adding Inbound Agent to Your Website
Copy Embed Code
From your agent's configuration (Step 3) or agent card menu, click "Copy Embed Code". The code is now in your clipboard.
Open Your Website Files
Access your website's HTML files or content management system (CMS).
Locate the Closing Body Tag
Find the closing body tag in your HTML. This is typically at the bottom of your page template.
Paste the Code
Paste the embed code just before the closing body tag. Save your changes.
Test on Your Website
Visit your website and verify the chat widget appears (usually bottom-right corner). Click to open and send a test message.
Integration Strategies
Full-Site Integration:
- Add embed code to your site template or master layout
- Chat widget appears on every page
- Best for: General customer support, FAQ assistance
Targeted Page Integration:
- Add embed code only to specific pages
- Examples: Product pages, pricing page, contact page
- Best for: Sales-focused agents, product-specific support
Multiple Agents:
- Deploy different agents on different pages
- Sales agent on pricing page, support agent on help center
- Technical agent on developer documentation
- Each agent trained with relevant knowledge for its page
Strategic Placement
Deploy sales-focused agents on high-intent pages (pricing, product demos, contact) and support-focused agents on help center and documentation pages for best results.
Knowledge Base Integration
How Agents Use Knowledge
Your Inbound Agent relies on its linked knowledge base to answer questions accurately:
- Agent retrieves relevant information from the knowledge base for each question
- Responses are generated based on your documentation, not generic AI knowledge
- Regular knowledge base updates ensure agents always have current information
- Well-organized knowledge bases produce better, more accurate responses
Updating Your Knowledge Base
Navigate to Knowledge Base
Go to AI SDR → Knowledge Base in the main navigation.
Find Your Agent's Knowledge Base
Locate the knowledge base associated with your Inbound Agent (created automatically during agent setup).
Add Content
Upload documents, add FAQs, paste product information, or input company policies.
Save Changes
Save your knowledge base updates. Your agent immediately uses the new content.
What to Include:
- Product features and capabilities
- Pricing information and plans
- Common questions and answers
- Company policies and procedures
- Integration and setup guides
- Use cases and best practices
Real-Time Updates
Knowledge base changes take effect immediately. Your agent will use updated information in new conversations right away—no need to retrain or reconfigure.
See Knowledge Base for detailed knowledge base management.
Analytics and Conversation History
Viewing Agent Performance
Access analytics at: AI SDR → Inbound Agent → Analytics Tab
Session History:
- View all conversations across all your agents
- See conversation start time, agent name, and status
- Click any session to view the full conversation transcript
- Pagination for browsing through conversation history
Filtering Options:
- By Agent: Filter conversations for specific agents
- By Date Range: Select "from" and "to" dates to narrow results
- By Status: Filter by conversation status (active, completed, etc.)
Statistical Aggregation:
- View conversation metrics over time
- Aggregate data by day, week, or month
- Select specific agents for focused analysis
- Choose your preferred timezone for reporting
Key Metrics
Track important indicators of agent performance:
Conversation Volume:
- Total number of conversations initiated
- Conversations per agent
- Peak conversation times
Engagement Metrics:
- Active sessions
- Average conversation length
- Response time
Trend Analysis:
- Conversation volume over time
- Day-of-week patterns
- Hour-of-day patterns
Analyze and Optimize
Regularly review conversation history to identify:
- Questions the agent struggles with (update knowledge base)
- Common visitor concerns (refine your messaging)
- Peak traffic times (ensure agent is always active)
- Opportunities to improve agent training
Use Cases and Best Practices
Use Case 1: Customer Support Automation
Goal: Reduce support team workload while providing 24/7 assistance
Agent Configuration
Role: Customer support assistant Knowledge Base: FAQ, troubleshooting guides, product documentation Deployment: Help center, support portal, all product pages
Expected Outcomes
• 40-60% reduction in simple support tickets • Instant responses to common questions • 24/7 availability for global customers • Faster time-to-resolution
Best Practices:
- Keep responses friendly and empathetic
- Train agent to escalate complex issues to human support
- Include links to detailed documentation
- Regularly review conversations to identify gaps in knowledge base
Use Case 2: Sales Lead Qualification
Goal: Automatically engage and qualify website visitors
Agent Configuration
Role: Sales assistant and lead qualifier Knowledge Base: Product benefits, pricing, case studies, ROI information Deployment: Pricing page, product demos, homepage
Expected Outcomes
• Capture contact info from 20-30% of engaged visitors • Qualify leads based on company size, use case, budget • Auto-route qualified leads to sales team • Faster lead response time
Best Practices:
- Ask qualifying questions naturally during conversation
- Collect email and company info for follow-up
- Integrate with Outreach Agent for automatic nurture sequences
- Provide clear next steps (book demo, download resources)
Use Case 3: Product Guidance and Onboarding
Goal: Help prospects understand features and use cases
Agent Configuration
Role: Product consultant and advisor Knowledge Base: Feature documentation, integration guides, use case examples Deployment: Product pages, feature comparison pages
Expected Outcomes
• Educate prospects on product capabilities • Recommend relevant features based on needs • Increase demo request conversion • Reduce pre-sales questions to sales team
Best Practices:
- Focus on benefits and outcomes, not just features
- Recommend specific features based on prospect's industry or role
- Provide comparison information if asked about competitors
- Offer product tours or demo scheduling
Optimization Best Practices
Agent Training
Be Specific About Role and Boundaries
The more specific your role prompt, the better your agent performs. Define exactly what it should do, how it should communicate, and crucially—what it should NOT do.
Provide Response Examples
Include examples of good responses in your role prompt. This helps the agent understand your preferred communication style and level of detail.
Iterate Based on Real Conversations
Don't guess—review actual conversations, identify issues, and refine your prompt accordingly. Most great agents are built through iteration.
Knowledge Base Management
Keep Content Current
Outdated information damages trust. Review and update your knowledge base monthly or whenever products, pricing, or policies change.
Organize for Retrieval
Structure knowledge base content clearly with headings, sections, and clear Q&A format. Well-organized content helps agents find accurate answers faster.
Cover Common Questions Thoroughly
Review analytics to identify frequently asked questions. Ensure these topics are covered comprehensively in your knowledge base.
Deployment Strategy
Match Agent to Page Purpose
Deploy sales-focused agents on commercial pages (pricing, product) and support-focused agents on help/documentation pages. Don't use one agent for everything.
Test Before Full Rollout
Deploy to a single page first, test thoroughly, refine based on conversations, then roll out site-wide.
Conversation Quality
Monitor Weekly
Review conversation transcripts weekly to identify:
- Questions the agent can't answer well (add to knowledge base)
- Opportunities to improve conversational flow
- Misunderstandings or errors in agent responses
Set Clear Escalation Paths
Train your agent to recognize when human help is needed and provide clear escalation instructions ("Let me connect you with our support team" + contact method).
Troubleshooting
Problem: Chat widget doesn't appear on website
Possible Causes:
- Embed code not correctly added to website
- JavaScript errors blocking widget loading
- Browser cache showing old version of page
Solutions:
- Verify embed code is pasted before the closing body tag
- Open browser console (F12) and check for JavaScript errors
- Clear browser cache and hard refresh (Ctrl+Shift+R or Cmd+Shift+R)
- Test in incognito/private browsing mode
- Confirm the script URL is accessible (not blocked by firewall)
Problem: Agent not responding to messages
Possible Causes:
- Agent status is set to Inactive
- Agent not fully configured (missing role prompt)
- Network connectivity issues
- Knowledge base is empty
Solutions:
- Check agent status toggle is set to Active
- Verify agent has a role prompt configured in Step 2
- Test agent in the built-in test interface (Step 3)
- Ensure knowledge base has content
- Check network connection
- Try refreshing the page
Problem: Agent provides inaccurate answers
Possible Causes:
- Knowledge base content is outdated or incomplete
- Role prompt doesn't clearly define response guidelines
- Information exists but isn't well-structured for retrieval
Solutions:
- Review and update knowledge base content
- Add specific information for topics where agent struggles
- Refine role prompt to be more specific about answer format
- Organize knowledge base with clear headings and structure
- Test with same questions after updates
Problem: Can't copy embed code
Possible Causes:
- Insufficient permissions (plan doesn't include Inbound Agent)
- Browser doesn't support clipboard operations
- Ad blocker or security extension blocking functionality
Solutions:
- Verify your plan includes Inbound Agent access
- Try a different browser
- Manually select and copy the code
- Disable browser extensions temporarily
- Contact support if issue persists
Problem: Auto-Enhance not working
Possible Causes:
- Role prompt field is empty
- Network connection issue
- Server temporarily unavailable
Solutions:
- Enter some initial prompt text before clicking Auto-Enhance
- Check your internet connection
- Refresh the page and try again
- Write your prompt manually if Auto-Enhance remains unavailable
Problem: Conversations are slow or laggy
Possible Causes:
- Large knowledge base taking time to search
- Very long role prompt
- Network latency
- High server load
Solutions:
- Optimize knowledge base—keep most relevant content
- Simplify role prompt to essential instructions only
- Check your internet connection speed
- Contact support if slowness persists across multiple tests
Next Steps
Ready to deploy your first Inbound Agent?
- Create your agent — Set name, voice, and role
- Train with knowledge — Upload documentation and FAQs to your knowledge base
- Test thoroughly — Use the built-in test interface to refine responses
- Deploy to your website — Copy embed code and add to your site
- Monitor and optimize — Review conversations regularly and improve over time
Never Miss a Lead Again
Inbound Agents work 24/7 to engage, qualify, and convert website visitors—even when your team is offline. Start providing instant, intelligent responses to every visitor today.
Related Resources:
- Knowledge Base — Train your agent with custom knowledge
- Outreach Agent — Automate follow-up with qualified leads
- AI SDR Overview — Explore the full AI SDR platform