Publish the Chatbot
🎉 Final Steps: Testing & Launch
You can either enable the site-wide chatbot option to display the chatbot on all WordPress pages of your site, or you can deploy it on specific pages using shortcodes.
There are two ways to deploy the chatbot on specific pages:
Option 1: Embedded Chat Interface
[ai_botkit_chat id="your-bot-id"]This shortcode embeds a chat interface directly within your page content.
Option 2: Floating Widget
[ai_botkit_widget id="your-bot-id"]This shortcode adds a floating chat bubble widget that appears in the corner of your site.
Simply copy either shortcode and paste it into any WordPress page or post where you want the chatbot to appear. The plugin provides a convenient copy-to-clipboard button for each option.
Sitewide Chatbot
You can use this option to add chatbot on all of your WordPress pages at once.
⚙️ Important Setup Requirements
Before Your Chatbot Will Work:
Activate Your Chatbot:
Make sure the "Activate Chatbot" toggle is enabled in Step 1 of the wizard.
Inactive chatbots won't appear on your website.
Configure API Keys:
Go to the Settings tab and add your AI provider API key (OpenAI, Anthropic, etc.).
Without valid API keys, the chatbot cannot generate responses.
Test Before Launch:
Use the Preview feature in Step 5 to test your chatbot's responses and appearance.
Make necessary adjustments before making it live on your site.
📊 Monitoring Performance
After Launch:
Check Analytics:
Conversations tab: See how many people are using your bot
Message history: Read actual conversations
Feedback scores: Monitor thumbs up/down ratings from users
Continuous Improvement:
Add new content: Keep knowledge base updated
Refine responses: Based on user feedback
Monitor costs: Check AI provider usage
Update settings: Adjust based on performance
🔧 Troubleshooting Common Issues
"Bot gives generic answers" Solutions:
Add more specific training content to the Knowledge Base
Increase context chunks in Step 4 (Model settings)
Upload more detailed documents or relevant URLs
Ensure your content is properly indexed by the system
"Responses are too short" Solutions:
Increase max tokens in Step 4 (Model settings)
Improve personality description in Step 1
Add more comprehensive training content
Adjust the AI model to a more capable version
"Bot doesn't match my brand" Solutions:
Adjust personality and tone in Step 1
Customize colors and fonts in Step 5
Update greeting messages in Step 3
Upload your brand logo as the chatbot avatar
"High AI costs" Solutions:
Use a smaller AI model in Step 4
Reduce max messages and context chunks
Lower max tokens setting
Review conversation volume and implement rate limiting if needed
"Knowledge Base not being used" Solutions:
Verify documents are properly uploaded and processed
Check that the chatbot is linked to the correct knowledge base
Increase the number of context chunks
Ensure content is relevant to common user queries
🎯 Success Tips
For Best Results:
Start Simple: Begin with basic functionality, expand later
Test Thoroughly: Spend time testing before going live
Monitor Regularly: Check conversations and feedback weekly
Keep Updated: Add new content as your business grows
Iterate Often: Continuously improve based on user interactions
Expected Timeline:
Initial Setup: 15-30 minutes
Testing Phase: 1-2 days
Launch: Immediate after testing
Optimization: Ongoing process
Knowledge Base Integration:
The quality and relevance of your training content directly affects the chatbot's ability to provide accurate, contextual answers
Regularly update your knowledge base as your products, services, or information changes
Monitor which documents are most frequently referenced and optimize them for better responses
Your AI assistant is ready to help your visitors 24/7! 🤖✨
Last updated