Analytics

The Analytics tab gives you a complete picture of how your companion is performing. All cards are displayed in a single continuous view, with a shared date range and granularity selector at the top. By default you see the last 30 days.

Use the granularity selector to switch between hourly, daily, weekly, or monthly groupings, and the date range picker to change the time period.

Layout

Row 1 — Six summary and quality cards in a 3-column grid:

• Query Resolution — resolved, partially resolved, and unresolved conversation breakdown.
• User Sentiment — positive, neutral, and negative sentiment split.
• Device Analytics — devices and operating systems your visitors use.
• Response Quality — a radar chart scoring your companion across five quality dimensions (relevance, correctness, completeness, clarity, tone fit) on a 1--5 scale, with an overall A--E letter grade.
• Audience Journey — lifecycle stage distribution showing where conversations fall (pre-sale, onboarding, active use, troubleshooting, renewal). Toggle between area and bar chart views.
• User Friction — a donut chart showing the percentage of sessions with friction, broken down by type (repeated questions, user corrections, clarification loops, task abandonment).

Row 2 — Three companion-specific analytics cards:

• Sales Signals — buying intent distribution (none, low, medium, high), conversion opportunity count, and actionable opportunities for the creator. This card only appears when your companion has conversations with detectable buying signals; otherwise it is hidden.
• Failure Analysis — a donut chart of failure root causes (knowledge gap, out of scope, hallucination, etc.) grouped by severity.
• Escalation Routing — the escalation rate with a breakdown of which teams (sales, support, technical, billing) receive escalated conversations. Color-coded by health: green, amber, or red.

Row 3 — Two cards side by side:

• Product Insights — product feedback and feature requests extracted from conversations.
• Recommended Actions — prioritized AI-generated suggestions grouped by owner (creator, sales, support, product).

Row 4 — Interaction trends chart and top keywords:

• Interaction Trends (wider) — conversation activity over time, groupable by intent, sentiment, or resolution.
• Top Keywords (narrower) — the most common topics visitors ask about.

Row 5 — Full-width geographic distribution:

• Session Locations — a world map with country heatmap and location leaderboard.

How the data is generated

After each conversation ends, Aura's AI analyzes the chat using a 10-step structured analysis. It classifies the context (intent, lifecycle stage), evaluates the outcome (resolution, sentiment), scores response quality across five dimensions, detects friction and sales signals, assesses any failures, determines escalation needs, extracts topics and product insights, and generates up to five recommended actions. This happens automatically.