Filter and explore analytics
Two flows in this article: filtering the dashboards down to what you care about, and using the AI features to interpret what the charts are showing you.
Filtering
Every view in Analytics responds to the same global filter controls. Setting a filter once applies it across every view you visit.
The filters
- Date range — Today, last 7/14/30/60/90 days, or custom dates
- Channel — Email, chat, voice, SMS, WhatsApp, Messenger
- Team — Filter to specific teams
- Country — Geographic breakdown
- Brand — For multi-brand organizations
Common filter patterns
- Last week’s performance — Date = Last 7 days
- Just my team — Team =
<your team> - Compare two teams — Filter to one, screenshot, then switch — Analytics doesn’t currently overlay two team filters in one view
- Just chat conversations — Channel =
Web Chat - Q1 retrospective — Date = Custom, Jan 1 – Mar 31
- Just one country’s customers — Country =
<country>
Filters persist across views
Set a filter on the Overview, then click into Agent Performance — the filter is still applied. Saves you from re-filtering at every view switch.
To clear filters, click Reset at the top of the filter bar.
AI Insights — get a narrative
Every chart in Analytics has an AI Insights button. Click it. Atender generates a narrative explanation of what the data shows.
Examples of what it produces:
“Response times increased 23% this week, primarily driven by a spike in email volume on Tuesday. The billing team was most affected.”
“CSAT improved in chat but declined in phone — this correlates with the new chat snippets deployed last Thursday.”
This is especially valuable for:
- Briefing stakeholders without manual analysis
- Spotting correlations you wouldn’t have noticed
- Sanity-checking your own interpretation of a chart
The narrative is regenerated each time you click — the AI is reading the current filtered data, not stored summaries.
Chat about this data — ask questions
Every chart also has a Chat about this data button that opens a floating AI panel. Ask natural-language questions:
- “Why did our response time spike yesterday?”
- “Which agent improved the most this month?”
- “How can I reduce our reopen rate?”
- “What’s causing the backlog growth trend?”
The AI responds with contextual, data-informed answers that reference your actual metrics. It’s like having a data analyst on call.
The chat is scoped to the chart you opened it from. To ask broader questions, navigate to the relevant chart first or use the custom analytics builder.
A typical investigation flow
- Start at Overview. Look at the trend indicators — what’s getting better or worse?
- If a KPI’s trending wrong, drill into its dedicated view. SLA dropping? Open Conversations & SLA. CSAT dropping? Open Customers & CSAT.
- Filter to the suspected cause: team, channel, time period.
- Click AI Insights on the chart that surprised you. Read the narrative.
- Click Chat about this data to ask follow-up questions: “is this driven by the new chatbot, or by staffing?”
- Drill further if the AI suggests a specific cause: e.g., open Agent Performance to see if a specific person is the source.
What the filters DON’T do
- They don’t change the underlying data — they only narrow what’s displayed.
- They don’t apply to AI Insights / Chat-about-data on a chart that’s already filtered — those features always read the filtered data on the chart they’re invoked from.
- They don’t persist across browser sessions — each new login resets to default. To save a frequent filter combination, use a custom analytics module instead (see Build custom analytics).