Skip to main content

Documentation Index

Fetch the complete documentation index at: https://setup.cevro.ai/llms.txt

Use this file to discover all available pages before exploring further.

Drivers is the main VoC surface inside Insights. It answers questions like:
  • “What are the top reasons players contacted us last month?”
  • “Which Themes have the highest escalation rate?”
  • “Which Needs are growing the fastest week-over-week?”
Every closed ticket is automatically classified into a three-level taxonomy (Theme → Need → Problem) and visualized here. URL: https://app.cevro.ai/insights/drivers

​
Top Bar

ControlPurpose
BrandNarrow to a single brand or “All Brands”
DatePreset ranges (last 7/30/60/90 days) or a custom range
Group bySwitch the aggregation level for every chart and the explorer table (Theme / Need / Problem)
Filter toggleReveals the advanced-filters panel

​
Group by

OptionBest for
Theme (default)High-level executive view
NeedDaily operations
ProblemGranular drill-downs

​
The Charts

Side-by-side bars comparing the current period vs the previous period for each group. The label above each bar shows percent change. Quick read on which Themes/Needs/Problems are growing or shrinking. Click a bar to drill into its tickets.

​
Top 10 contact drivers

Horizontal bar chart of the top 10 groups by ticket volume. Click any bar to drill in.

​
Top escalation drivers

Top 10 groups ranked by escalation rate. Useful for finding pain points the AI is handing off most often.

​
Top resolutions

Top 10 groups by resolution rate — the flip side of escalation, where your AI is winning.

​
Volume trend

A line chart of ticket volume for the top groups across the selected window. Lets you spot momentum changes that aren’t yet visible in the period-over-period bars.

​
Theme volume treemap

Each rectangle is a Theme/Need/Problem; size = volume, color = average CSAT (red <2.5 → amber → light green → green ≥4.0). Click any tile to drill in.

​
Topic outcome

Stacked horizontal bars showing for each group: how many tickets were Escalated, handed off via Delegated Resolution, or Resolved by the AI. The fastest way to see where AI is winning vs. handing off.

​
Topics Explorer table

The most powerful component on the page. A hierarchical table you can expand from Theme → Need → Problem.
ColumnDescription
Area / Topic / IssueThe Theme/Need/Problem, indented by hierarchy
VolConversation count
%Share of total
AHTAverage Handle Time
Esc%Escalation rate
Del%Delegation rate
Res%Resolution rate
SentimentNegative-sentiment share with a colored chip
TrendSparkline across the selected window
Every numeric column is sortable. The View conversations button at the end of each row opens the Conversations Modal scoped to that node. Use the Expand all button above the table to fan out every Theme → Need → Problem in one click.

​
Drill-down: Conversations Modal

Almost every chart and every “View” button opens the Conversations Modal — a paginated list of the underlying tickets. Each row shows ticket ID, Title, Brand, and Creation Time. Click a row to open the full conversation in a new tab. The modal inherits the current page-level filters (date, brand, taxonomy node, advanced filters). It’s the bridge between aggregated insight and ground-truth conversation — use it freely, quick drill-downs are the fastest way to validate a number.

​
Advanced Filters

Click the Filter toggle for the full filter panel. Available filters:
  • Signals — tickets that triggered one or more named signals
  • Contact drivers — pick specific Need-level drivers
  • Status — Resolved / Escalated / Delegated
  • Agent — narrow to a specific AI agent
  • Browser, Platform, Language, Channel (Web / Email / WhatsApp / Messenger / SMS / API)
  • Polarity — Positive / Neutral / Negative tone
  • State — Urgent / Confused
  • Contact type — Complaint / Help-or-question
  • CSAT Rating — 1★ … 5★
  • Scorecard — score-range filters per section (when QA Scoring is enabled)
  • Page — substring match on the page URL where the chat started
  • Player metadata — dynamic conditions on contact attributes (e.g. country = Brazil, lifetime deposits > $10k)
The badge on the filter toggle counts how many advanced filters are active.

​
Tips

  • Start at Theme, drill to Problem. Begin with Group by Theme for the executive view, then drop into Need / Problem when you need to act.
  • Compare windows. The Trending chart and the Trend sparkline make week-over-week shifts obvious. Don’t over-react to single-day spikes.
  • Cross-reference with Sentinel. A spike in a Theme + low scorecard quality on those same tickets = an actionable training brief.
  • If a Theme has no Need-level rows you recognize, head to Taxonomy and review the Discovery suggestions — emerging patterns may not be in your taxonomy yet.
  • Taxonomy — the catalog of Themes, Needs, and Problems
  • Feedback — CSAT clustering, the service-quality counterpart to VoC
  • Signals — message-level detection (the “what is happening during the conversation” view)
  • Overview — the executive landing page that pulls from this data