WATI chatbot analytics

WATI counts messages sent and delivered. This counts whether the conversation worked, on the same WhatsApp traffic, without changing WATI.

WATI's chatbot analytics reports sessions and what happened to them: completed, dropped off, reassigned, in progress, and the three rates derived from those.

Read the definition of completed carefully. It means the agent reached the end of the flow. That is a fact about the flow, not about the customer.

Where the two readings separate

WATI says completed	The agent walked to the end of its flow.
WATI says dropped off	The contact stopped responding, timed out, or something broke.
We say resolved	The customer got the thing they asked for, judged against your rules.

An agent that answers the wrong question confidently and reaches the end of its flow is a completed session and a failed conversation. Both are true at once.

WATI's AI agent dashboard does report a resolution rate, defined as conversations resolved without human assistance. That is containment: it counts who handled the conversation, not whether the answer was right.

What we read instead

Whether the agent answered the question that was asked.
Whether it said it was not a person when asked.
Whether it promised a date or a price nobody authorised.
Where the customer stopped replying, and why.

The reasons customers stopped are grouped and ranked by how many conversations each one cost, so the list is ordered by what to fix first.

Connect it

Point WATI's webhook at the Evidova URL from Setup.
We read the same messages WATI already posts. Nothing in your inbox changes.
Approve the rules for your vertical.
Scores start on the traffic that arrives next, and on any history you import.

WATI does not sign what it sends, so the token in the webhook path is what protects it. Treat it like a password.