Once you launch an AI Chatbot, it’s easy to watch the conversations roll in and call it a day. Then the novelty wears off, and a bigger question comes up:
Is my chatbot actually making an impact?
To help you answer that, we’ve recently added an Analytics section to the dashboard. It includes key metrics such as chat sessions, messages received, AI replies sent, AI-handled chat time, and leads collected.
- Leads / Chat sessions → What percentage of conversations result in a lead?
If this ratio is lower than you expected, review conversations where visitors showed interest but didn’t leave their details. You may find that the lead request appears too early, asks for too much information, or doesn’t clearly explain what happens next.
- Messages / Chat sessions → How much interaction happens during an average conversation?
A low number may mean visitors get an answer quickly, but it could also mean they leave after the first response. A high number could signal strong engagement - or that visitors need several attempts to get a useful answer. Reviewing a few short and long conversations will help you tell the difference.
- AI-handled time / Chat sessions → How long does the chatbot handle an average conversation?
An increase may mean the chatbot is taking on more substantial questions. But longer conversations aren’t always better, so check whether they end with a useful answer, a lead, or a clear next step.
The key is not to optimize every number at once. Start with the goal that matters most to you:
- If your chatbot is meant to generate leads, focus on leads per session.
- If it’s meant to answer questions, look at engagement alongside conversation quality.
- If it’s meant to reduce repetitive work, examine which questions the chatbot handles and whether visitors still need help afterward.
Say two chatbots both collect 500 leads. One gets them from 5,000 sessions, while the other needs 50,000 sessions. The total is the same, but their lead conversion rates are 10% and 1%—two very different results.
Analytics tells you where to look. The conversations themselves help you understand why the numbers look that way.
What’s your chatbot’s main goal, and which metric would you use to measure it?