​

Analytics

The Analytics section shows how your agents are performing: how many conversations they get, how many the AI resolves on its own, how many go to a person and how long your team takes to reply.

• All agents: side menu → Analytics.
• A single agent: My agents → [your agent] → Analytics.

​

Top bar

At the top of the page you have:

• Tabs: Chats, AI performance, Human support and Reviews.
• Agent (overview only): All agents or a specific one.
• Period: Last 7, 14, 30 (default), 60 or 90 days.
• Grouping: Daily or Weekly, for the charts.

Many metrics are compared with the previous period of the same length (for example, the previous 30 days).

​

Chats tab

​

Volume

Below, the Conversations chart compares this period with the previous one and shows the peak and the daily average.

​

Distribution

• Channels — conversations by channel (Web, WhatsApp, Instagram, Telegram, Messenger, Email, Google Reviews, API).
• Conversation status — how many are Open, Resolved and Closed.
• When they chat the most — heat map of visitor messages by day of the week and hour (always for the last 90 days). It helps you know when to have more people available to handle handoffs.

​

Audience

• Top countries — where conversations come from.
• Conversation starting pages — which pages of your site chats are opened on.
• Captured leads — the period’s leads with name, email, phone, channel, agent and date. View all takes you to the Leads section.

​

Agent comparison

In the All agents view, the Chats tab adds a table with one row per agent: conversations, deflection (percentage resolved without a person), escalations, credits consumed and a trend line.

​

AI performance tab

​

Autonomous resolution

• Deflection rate — percentage of conversations the AI resolved without needing a person.
• Response satisfaction — percentage of replies visitors marked as Helpful with the widget’s rating buttons, and how many received a rating.

​

Escalation to humans

• Escalation rate — percentage of conversations that went to a person.
• Escalated conversations — how many there were in the period.
• Main reason — still under construction.
• AI credit usage — credits used by the agent’s replies and the average per conversation.

The Credits consumed per day chart includes a projection for the month and the percentage of your plan it represents. For more detail on usage, see AI models and credits.

​

Human support tab

Measures your team’s handling of escalated conversations (handoff). Times are counted from when the visitor asks to talk to a person. Each metric shows the median and the p90 (the time that 90 % of cases don’t exceed).

• First reply by day — chart with a target line. In a single agent’s view, the target is the first reply time you set in its Human Handoff Action; if there isn’t one, 3 minutes is used.
• Load per human agent — how many handoffs each person on your team handled and their reply time.

If no conversation went to a person during the period, the tab says so (“No visitor needed a human”) along with the percentage the agent resolved. In a single agent’s view, the Configure handoff button also appears.

​

Reviews tab

Shows the Google reviews of the locations connected through the Google Reviews channel. Below are Reviews per day and Average rating per day, the Star distribution, the Sentiment (positive, neutral or negative), the Most mentioned topics and the By location table with reviews, rating, answered and to approve for each location.

​

Frequently asked questions