Insights overview

Pulse, Menu, Waiter, and Customers, plus date ranges, branch filters, and CSV export.


Insights turns every scan and question into numbers you can act on. It is organized into four tabs, each answering a different question about your restaurant.

The Insights dashboard on the Pulse tab, with four headline metrics across the top1234
Insights, Pulse tab. Four headline metrics, each with a trend against the previous period.
  1. Conversations, total chats with the AI waiter.
  2. Answer success, share resolved without falling back.
  3. Customer mood, average sentiment.
  4. Open actions, opportunities you have not triaged.

The four tabs

Pulsehealth
The overall snapshot: volume, answer quality, customer mood, and open actions.
Menudishes
How individual dishes perform, what gets attention, and what gets ignored.
WaiterAI quality
How well the AI waiter answers, and where it falls back or escalates.
Customerspeople
Who is asking, in what language, how they move through the visit, and how they feel.
Moving between the four Insights tabs: Pulse, Menu, Waiter, and Customers.

Filter and export

Every tab shares the same controls in the header.

  1. 1
    Pick a date range

    Choose a preset or a custom range: Today Yesterday Last 7 days Last 30 days Last 90 days Custom range. Trends compare against the previous window of the same length.

  2. 2
    Scope to a branch

    If you run more than one location, use the branch filter to focus on one, or compare them on Pulse. Single-location restaurants do not see this control.

  3. 3
    Export to CSV

    Select Export CSV to download the current view. Large exports are capped; if that happens, narrow the date range for a complete file.

  4. 4
    Open the action center

    Select Actions to open the opportunities drawer. See opportunities →.

Pulse

The health snapshot. Four headline metrics sit at the top, each with a trend against the previous period.

Conversationscount
Total customer chats with the AI waiter in the range.
Answer successpercent
Share of conversations resolved without falling back or escalating.
Customer moodsentiment
Average sentiment graded across the chats we could read.
Open actionscount
Opportunities you have not yet resolved or dismissed.

Below the metrics: a conversations-by-hour chart, an hourly volume heatmap across the week in your venue timezone, a preview of open opportunities, and, for multi-location restaurants, a branch comparison.

The Menu tab of Insights, showing recommendation impressions, taps, and dish performance
The Menu tab. How dishes perform as recommendations, and which ones need a push.
Impressionscount
How often dishes were shown to customers as AI-waiter recommendations.
Tapscount
How often customers tapped a recommendation to open the dish.
Click-through ratepercent
Taps divided by impressions.

The menu map plots every dish by visibility against conversion, sorting them into stars, hidden gems, and dishes that need a push. Each dish links straight to its editor so you can improve the listing.

Waiter

The Waiter tab of Insights, showing answer success, fallback, escalation, and response time
The Waiter tab. Answer quality over time, and the outcome of every conversation.
Answer successpercent
Conversations resolved without falling back or escalating. Higher is better.
Fallbackpercent
Conversations where the waiter could not answer. Lower is better.
Escalationpercent
Conversations handed to a human. Lower is better.
Response timeseconds
Average time the waiter took to reply.
Fallbacks are your to-do list

A rising fallback rate points at a teaching gap. Open the Waiter tab to see the most common reasons, then fix them on the dish or in the waiter's house rules. See teach your waiter →.

Customers

The Customers tab of Insights, showing the visit funnel, sentiment, and top questions
The Customers tab. The visit funnel, language mix, mood, and most-asked questions.

The Customers tab follows the visit as a funnel, from customers who reached the menu, to those who opened a chat, to those who asked a question. It also breaks down languages, tracks mood over time, and lists the top customer questions.

FAQ

What does 'customer mood' actually measure?

An automatic sentiment grade, averaged across every conversation we could read in the range. It shows a placeholder when there is not enough data yet.

Why don't I see a branch filter?

The branch filter only appears when your organization has more than one location.

My CSV export looks cut off.

Very large exports are capped. Narrow the date range and export again for a complete file.

Do 'Today' and 'Yesterday' stay fixed if I bookmark them?

No. They recalculate on each load, so a bookmarked link always shows the current today or yesterday.

Was this page helpful?

Last reviewed July 16, 2026