Ticket details, sentiment and sanity scores

How to read the ticket details dialog, the sentiment analysis dialog and the sanity score that you open from the ticket list on Home.

Written By Chris Scaminaci

Last updated About 3 hours ago

Three windows on Home show what QuantumOps worked out about one ticket without leaving QuantumOps: the ticket details dialog, the sentiment analysis dialog and the sanity details. This page explains what each one shows and how to read the scores in them.

The ticket details dialog is not the Ticket Dashboard. The Ticket Dashboard opens only inside HaloPSA. The dialog on this page opens from QuantumOps.

Before you start: any signed-in user can open these dialogs. The ticket must have been analysed. For a ticket that was not, the AI Analysis section says This ticket was not processed with AI analysis. How ticket analysis works explains which tickets are analysed.

Open the ticket details dialog

On Home, find the ticket in the list. Use Search or the quick filters if needed.
Select View Details, the eye button at the end of the row. Double-clicking the row does the same.
The dialog opens with the title Ticket #N - Subject: followed by the start of the subject.

You can also open it from a list behind a tile or a workload count, with the eye icon (QuantumOps details). See Open the tickets behind a number.

Read the header

The top of the dialog is a row of badges.

Badge	What it shows
Client	The client. Select it to open Client Profile, the profile QuantumOps built for that client, when one exists. See The Client Dashboard.
User	The person who raised the ticket. Select it to open User Profile, when one exists. See The User Dashboard.
Actions	How many actions the ticket has in HaloPSA.
Health	The health score as a percentage, or a dash when the ticket has never been assessed. Red below 60, orange from 60 to 79, green from 80. Health starts at 100 and drops for each problem found: an SLA breach, inactivity, an overdue agent follow-up, a ticket older than the maximum age, a late assignment or a long SLA hold. See How the health score is calculated. Each screen bands the score differently; see Where the score shows.
Sanity	The sanity score as a percentage. Select it to open Sanity Analysis. A grey 0% means the ticket has no usable score. See Understand the sanity score.
Status	Open, Closed or Unknown.
Date Created	When the ticket was created, in your organisation's time zone.
Last Communication	When the ticket was last updated.
Last Analysis	When the analysis was last refreshed, and how many times it has been (3x). A dash means it has not been refreshed.
Category	The ticket's category.

Next to the title, small badges mark the alerts the ticket has: Sentiment, Tonality, Alert and Suppressed. Point at Sentiment or Tonality to read the score and category.

Read the sections

Below the header are collapsible sections. Select a section's title to collapse it.

AI Analysis is the AI's summary of the ticket, as numbered cards. A card can carry a small tag such as Recommendation, Action, Issue or Root Cause.
Ticket Timeline lists what happened to the ticket, newest first.
Stale Review appears only when the ticket is stale. Issues Identified says why it went quiet and Recommended Actions says how to restart it. See Stale ticket detection.

At the bottom, Open in Halo opens the ticket in HaloPSA and Reprocess Ticket runs the full analysis again, except for a ticket that was fully analysed in the last two hours. Reprocessing spends AI tokens.

The ticket timeline

Each event shows a title, the time and, when the summary changed, what changed. The event titles are Ticket Closed, Stale Check Run, End User Email, End User Update, Agent Update and AI Incremental Refresh. Small tags on an event show the sanity and health scores at that moment, a Stale tag, and Trajectory Updated when the resolution paths changed.

Repeated Stale Check Run and AI Incremental Refresh events fold into one line such as AI Incremental Refresh × 4. Select the line to see each event. Projected Next Steps lists the steps of the first resolution path QuantumOps proposed for the ticket. The timeline updates by itself while the dialog is open.

Read the sentiment analysis dialog

Select View Sentiment Analysis on a row to open Sentiment Analysis - Ticket #N. The button is greyed out until the ticket has been scored.

The dialog shows:

Client, User and the ticket Summary.
Sentiment Score and Tonality Score, each out of 10, with a category under the number. Sentiment describes how the customer comes across, from Hostile and Frustrated through Neutral to Cooperative and Pleased. Tonality describes how professional the exchange is, from Abusive and Unprofessional through Neutral to Professional and Excellent.
Sentiment Alert Active and Tonality Alert Active when the ticket currently has those alerts.
Severity and Risk Category, when there is something to report.
Analysis, the AI's reasoning, then Key Phrases, Warning Indicators and Recommendations.
ESCALATION REQUIRED, with the kind of escalation (for example Manager, CSM or Immediate), when the AI judges the conversation needs someone senior.

Open in Halo opens the ticket in HaloPSA.

A score of 1 is the worst and 10 the best. An alert appears when a score is at or below the alert threshold. Administrators set the thresholds in Q-Director under Classification, in Sentiment & Tonality Analysis, and can shift them for one client or one contact. See Sentiment and tonality alerts and Q-Director: sentiment and categorisation.

Understand the sanity score

The sanity score tells you how closely two AI analyses of the same ticket agree. QuantumOps analyses a ticket when it processes it, and asks for a second analysis when it runs a sanity check. It then compares four things and scores the match out of 100.

A low score flags an uncertain analysis, not a bad ticket. The comparison uses the two analyses only. Neither is checked against the original ticket. A high score means the two analyses agree. It does not prove they are right.

Where the score shows

Where	How it reads the score
Sanity column on Home and the Sanity badge here	Red below 75, yellow or orange from 75 to 89, green from 90. A dash or a grey 0% means no score.
Sanity Score tile on Home	The average over the last 30 days. Healthy from 75, Moderate from 50, Needs Review below that.

Read the sanity details

In the ticket details dialog, select the Sanity badge to open Sanity Analysis. A ticket that was never checked has nothing to open. A ticket that was checked but gives a grey 0% opens the dialog, which then explains that nothing could be compared.
Analysis lists the findings. Key discrepancies: names each comparison that differed. Overall agreement: is Strong from 90, Moderate from 70, Weak from 50 and Contradictory below 50.
Metrics is a table with one row per comparison. Primary is the first analysis, Secondary is the second, Agreement says how well they match and Impact is the points taken off.

Row	What is compared	Impact when they differ
Service type	The kind of service the ticket is judged to be	20 points
Category match	Whether the ticket's category fits the issue (true or false)	20 points
Summary overlap	How many terms the two summaries share	5 points for a moderate match, 10 for a weak one
Tonality overlap	How many terms the two tone analyses share	5 points for a moderate match, 10 for a weak one

The score is 100 minus the points taken off. A row whose value is missing on either side is skipped. When nothing can be compared, the ticket has no score. The details then say Sanity details unavailable for this ticket. or that no comparable pair exists. Older tickets can show an earlier layout.

Run the check again

Select Sanity Check on the ticket's row on Home. QuantumOps asks the AI for a fresh second analysis, compares it with the first and updates the score. Then open the sanity details to read the result. Reprocess Ticket runs the check too, but only when a full analysis actually runs: a ticket fully analysed in the last two hours is not analysed again (see Home: the QA Pipeline). Re-Process Ticket on the Ticket Dashboard in HaloPSA follows the same two-hour rule; see Use the action buttons. Sanity Check runs whenever you select it.

Each sanity check spends AI tokens from your organisation's allowance. A lighter refresh that follows a ticket update can also move the score. Run Sanity Check when you want it recalculated from two fresh analyses.

Was this helpful?

Still need help? Ask the team