Analytics & AI Insights

The dashboard turns raw conversations into trends and AI-generated summaries so you can see how your agent is performing.

Analytics

Open the Analytics tab. Pick a time range — 7, 30, or 90 days — with the pills at the top. You'll see:

KPIs — totals for conversations, visitor messages, bot replies, leads, bookings, and handoffs.
Daily timeseries — how each of those metrics moves day by day.
Language breakdown — which languages your visitors use.
Hourly distribution — what time of day your widget is busiest.

Charts adapt to light/dark theme automatically. KPIs and charts are free on every plan.

AI Insights (weekly summary)

Below the charts, BotChata generates a short weekly summary of the last 7 days of conversations:

Overall sentiment of your visitors.
Top topics — what people asked about most, with a count and an example.
Observations — notable patterns worth acting on.

The summary is written in your dashboard language (the verbatim examples keep their original language). It refreshes in the background and needs at least 5 conversations in the period to generate.

Weekly AI Insights are available on Business and above. On lower plans you'll see an "Upgrade" card here instead.

Detailed AI Insights

The separate AI Insights tab goes deeper (Premium and above). For a 7- or 30-day range it clusters conversations into themes grouped by sentiment, each with:

Sub-groups and counts,
A representative example,
A concrete recommendation you can act on,
Links to the actual conversations behind the theme — click a chip to read the transcript.

Like the weekly block, it caches for ~7 days and refreshes in the background, and needs at least 5 conversations.

Chat events in your own Google Analytics

If your site already runs Google Analytics 4 or Google Tag Manager, the widget reports two events to it. It loads no tracker of its own and does nothing on a page without one.

botchata_chat_started — a visitor started a new conversation.
botchata_product_click — a visitor clicked a product card in the chat; the product's name is in the product_name parameter.

With GA4 installed directly (the gtag.js snippet), the events show up in Reports → Engagement → Events within a day. To count a started chat as a conversion, mark botchata_chat_started as a key event in Admin → Events. To see product names in reports, add product_name as an event-scoped custom dimension in Admin → Custom definitions.

With Google Tag Manager, the events arrive in the data layer. Create a Custom Event trigger with the event name botchata_chat_started (and one for botchata_product_click), then a GA4 Event tag that fires on it. To send the product name, add a Data Layer Variable product_name and pass it as the event parameter product_name on that tag. If the page also loads GA4 through gtag.js, the events already reach GA4 directly — don't add the GTM tag, or each one is counted twice.

The widget does not add UTM tags to links on your own site, so your ad campaigns keep the credit for the visits they bring.

Tips

Watch the top topics to decide what content to add next — recurring questions are FAQ candidates.
A rising share of one language is a signal to translate more of your site's content.
Use the hourly distribution to schedule handoff working hours when your team is actually around.