Bot Dashboard & Metrics
What every card on your bot's Dashboard tab means — satisfaction, resolution rate, the performance chart, top questions — and what to do about the numbers.
On this page
Every bot has its own Dashboard tab — the first thing you see when you open a bot. It answers two questions: is the bot doing a good job, and how much is it being used? This article explains what each card measures and, more usefully, what to do when a number looks off.
The two outcome metrics
The two large cards at the top are the numbers you're accountable for.
Satisfaction Rate
The average of user feedback ratings across conversations. Positive ratings count as 100%, neutral as 50%, and negative as 0%.
Feedback comes from the thumbs up / thumbs down control on bot answers — so this metric only moves if Collect Feedback is enabled in your widget customization (see Customizing the Chat Widget).
If satisfaction drops, open the Activity tab and filter to Negative conversations to see exactly which answers people disliked — see Reviewing Conversations.
Resolution Rate
The percentage of conversations marked as resolved — meaning users found their answer without needing a human.
A low resolution rate usually means one of two things: the bot lacks training on the topics people actually ask about, or people are asking things a bot can't do (refunds, account changes). The Top Questions card below helps you tell the difference.
The volume strip
Below the outcome cards, three smaller cards give volume context:
Metric | What it counts |
|---|---|
Interactions | All-time billable interactions — one interaction is one AI bot reply, across customer chat and Help Scout Assist. This is the metered unit your plan allowance is measured in. |
Messages received | Every message users have sent, including follow-ups. |
Messages answered | Every response your bot has generated, with the response rate — the percentage of conversations that got at least one bot reply. |
Messages received will normally be higher than interactions: user messages, greetings, and errors aren't billed. For the full billing rules, see Understanding Interactions.
The Performance chart
The Performance card plots daily data as a line chart. Use the toggle in the card's top-right corner to switch between two views:
- Interactions — daily interaction volume. Spot traffic spikes, quiet periods, and the effect of deploying to a new page or site.
- Satisfaction — daily satisfaction rating on a 0–100% scale.
The most useful pattern to watch for: a volume spike paired with a satisfaction dip. That usually means a new group of visitors is asking about something the bot isn't trained on yet.
If the chart shows "No Performance Data," the bot hasn't had any interactions yet — data appears once real conversations start flowing in.
Top Questions
The Top Questions card lists your most frequently asked questions, each with:
- How many times it was asked (e.g. "14 times").
- A satisfaction bar showing how happy people were with the answer.
This card is your training to-do list. Read it two ways:
- High count, low satisfaction — lots of people ask this and the bot answers it badly. Fix the answer first: add or improve a training source that covers it, then verify in Train & Test Bot. See Training Sources Overview.
- High count, high satisfaction — a popular question the bot handles well. Consider adding it to your Suggested Questions so new visitors find it faster.
Recent Activity
The Recent Activity card shows the latest conversations: who asked (with a country flag when location is known), how long ago, an AI-generated one-line summary, and satisfaction / resolution badges where they matter.
Rows that need your attention are highlighted — negative-satisfaction conversations get a red edge, and conversations that have sat unresolved for more than a day get an amber one.
Click any row (or View all) to jump to the Activity tab, where you can read the full message thread, filter, and replay conversations. That workflow is covered in Reviewing Conversations.
Common questions
Why is my Satisfaction Rate stuck at 0% (or blank)?
Most likely nobody has left feedback yet, or Collect Feedback is turned off in your widget settings. Ratings are optional — many visitors never click thumbs up or down, so expect this metric to move slowly on low-traffic bots.
Interactions here doesn't match my Usage page — why?
The Dashboard tab shows this bot's all-time interactions. The sidebar Usage page shows the current billing period across your whole organization. See Tracking Your Usage.
The dashboard is empty. Is something broken?
If the bot is newly deployed, probably not — every card has an empty state until real conversations happen. Test it yourself from the Train & Test Bot tab (test conversations appear in your data), or check the deployment troubleshooting steps if you expected traffic that isn't arriving.
Where to go next
- Reviewing Conversations — read the actual conversations behind these numbers.
- Train & Test Mode — try fixes before your visitors see them.
- Tracking Your Usage — see how your interactions count against your plan.
Was this article helpful?