Using the Insights Dashboard

The Insights tab surfaces AI-generated patterns across all evaluated interactions. Unlike Metrics (which shows aggregate numbers), Insights tells you what is happening and why.

Insights can be filtered by agent (via dropdown) and date range (same presets as the Audit tab). When you select a relative date range (such as "Last 7 days"), it is saved by name so it stays relative when you return to the tab -- the range always reflects the current date, not the date you first applied it. A Refresh button on the Insights tab lets you reload the data manually at any time.

7.1 Five Insight Categories

Agent Performance:

Category	What It Surfaces
Top Strengths	Positive agent behaviors identified across interactions
Opportunities	Areas needing improvement, with coaching recommendations

Product Insights:

Category	What It Surfaces
Product Bugs	Technical issues surfaced from customer conversations
Product Feedback	Feature requests and suggestions mentioned by customers

Competitor Insights:

Category	What It Surfaces
Competitors	References to competitors in customer calls

Product Bugs, Product Feedback, and Competitors are all generated by a single configurable agent in the standard template, not three separate mechanisms -- they're grouped here by what they surface, not by how they're produced.

7.2 Each Insight Contains

Each insight category row shows an AI-generated summary at the top before the individual instances. Summaries are generated on demand as rows scroll into view -- only visible rows fetch a summary, so rows off screen cost nothing. If a row's summary cannot be generated, that row shows "Summary unavailable" while the other rows continue to load normally.

Column	Description
Driver	High-level behavioral topic with instance count (e.g., "Information verification")
Category	Sub-category within the driver (e.g., "Active listening")
Summary	AI-generated summary of the pattern across all instances in that row, loaded lazily as the row comes into view.
Instances	A button showing the total count that opens the detail drawer

7.3 Insight Thresholds

Not all insights are displayed. The system applies minimum thresholds to surface only meaningful patterns:

Agent Performance insights (Strengths and Opportunities):

Scope	Minimum Count Required
Single agent	>= 2 instances
Team (all agents)	>= 3 instances

Product insights (Bugs and Feedback): Shown if the insight is in the top 20% by count OR has a count >= 5.

Competitor Mentions: All mentions are displayed (count >= 1).

7.4 Viewing Insight Details

Click the Instances button on any insight row to open a detail drawer showing:

Detail	Description
Summary	The AI-generated summary for that row. If the row's summary was already loaded in the table, the drawer shows it immediately with no additional request.
Matching instances	Individual instance cards, each showing the UCID (linked to the Audit tab), justification, and source verbatims in an expandable accordion

7.5 Using Insights Effectively

Category	How to Use It
Top Strengths	Share with teams for positive reinforcement. Document best practices.
Opportunities	Build targeted coaching plans. Prioritize by count -- higher counts = more widespread issues.
Product Bugs	Route to engineering and product teams. These are real customer pain points.
Product Feedback	Forward to product management as voice-of-customer data.
Competitors	Share with strategy and marketing. Track which competitors are mentioned most.

7.6 Coaching Reports

Generate a personalized training document for an individual agent, tailored to their specific improvement areas:

From the Insights tab, select the agent to generate a report for
Ema analyzes that agent's QA feedback and insights
A training document is produced, focused on that agent's specific improvement areas

7.7 Ask Ema

Ask Ema lets you type a question about your QA data in plain English and get an answer built from the real numbers. It appears as its own tab on the AI Employee, next to Audit, Metrics, and Insights.

Ask Ema reads your existing QA results. It never grades a call, changes a rule, or edits your data -- every question becomes a read-only lookup.

Asking a Question

Type your question in the box at the bottom of the tab. Enter sends it; Shift+Enter starts a new line. Ema shows its progress while it works, then the answer appears in the conversation, usually with a chart or table alongside the text.

An empty chat offers three starting points:

Suggestion	Asks
Agent scores	Which agents had the lowest scores last week?
Rule failures	Which rules fail most often for the team?
My performance	How did I do this week?

What You Can Ask

Ask Ema reaches the QA data this AI Employee has produced -- scores and evaluations, rules and parameters, agent performance, CSAT, insights, transcripts, reviewer feedback, and imports.

Ask in business language, not database language. "Which rule fails most often on escalated calls?" works; you never need to know a table or column name.

Conversations and History

Conversations are kept, and Ema remembers earlier turns within one, so you can follow up with "and the month before?" without repeating yourself. New chat starts fresh -- use it when you change subject. Show history lists past conversations, which you can Rename or Delete.

What You Can See

Ask Ema never shows more than your role already allows elsewhere in Agent QA. Access is checked before your question is looked at, and the limits are applied to the lookup itself rather than filtered out afterwards.

Limits and Troubleshooting

Limit or problem	Detail
Read-only	Ask Ema can only look. It cannot re-run an evaluation, change a rule, or delete anything.
Scope	One AI Employee at a time. Ask Ema does not compare across AI Employees.
"I couldn't answer that"	Rephrase more specifically -- name the timeframe and what you want counted.
No Ask Ema tab	Ask Ema is not switched on for your organisation, or your role has no metrics access.