Hello, hope you can assist measuring our chatbot effectiveness.
We're creating an automated flow for our customer call center. At the end of the automated flow we'd like to ask the customer a question (Q) with buttons (B) to measure the effectiveness of the bot:
Q: Are you helped sufficiently?
B1: Yes, I'm helped sufficiently.
B2: No, speak to an agent.
If B1 is selected, the chat is closed and we start the feedback flow.
We interpreter this as a successful chatbot conversation
If B2 is selected, we handover to an agent.
We interpreter this as a unsuccessful chatbot conversation.
If the customer doesn't reply or decides to type another question in chat directly, we interpreter this as a neutral conversation.
The issue with this is that the customer has no incentive to click B1, and I just don't think it's the best possible solution.
How do you measure how effective your chatbot (automated) conversations are?
what are the best practices?
Thanks!
Thomas
Comments
2 comments