AI Chatbot Analytics: What to Track Before You Sell It

A chatbot demo without analytics is a guess

If you are showing AI chatbot demos to potential clients, you need to know what they actually do:

• Which demo they opened
• Which prompt they clicked
• What kind of message they sent
• Whether the AI replied
• Whether they clicked WhatsApp
• Which UTM campaign brought them in

My event taxonomy

For the portfolio chatbot demos, I track:

| Event | Meaning | |---|---| | page_view | Visitor opened a page | | ui_click | Visitor clicked a tracked UI element | | whatsapp_click | Visitor clicked a WhatsApp CTA | | ai_demo_message_sent | Visitor sent a chatbot message | | ai_demo_response_received | The AI route returned a response | | lead_form_submit | Visitor submitted a project brief |

Firebase + Clarity

Firebase stores structured events and lead records. Microsoft Clarity adds behavior context: session recordings, heatmaps, scroll depth, and interaction issues.

Why this matters

Analytics turns a portfolio from a static website into a sales system. You can see which service pages create demand, which projects create trust, and which demos convert into WhatsApp conversations.