The queue at a glance
Volume, response and resolution times over any period, by product and channel.
building a spreadsheet every MondaySupport · Insights
See the whole queue.And who did the work.
Volume, response and resolution times, how much Tore AI resolved, how good each conversation was, and which customers are drifting away while there's still time to help.
Tore opens soon. Join the waitlist and we'll email you the day you can start.
app.tore.ai/northwind/insights
Example data Example insights dashboard for Northwind over 14 days, then filtered to Northwind Mobile. It shows daily conversation volume with a spike after a release, who resolved conversations (Tore AI, AI-assisted, or a person, with agent hours saved), median first response against target, median resolution time, and at-risk customers: Beacon Data is high risk because of rising escalations, sentiment drift and churn-intent language, with a button to open the account.
No spreadsheet.No 2% sample. No surprises.
The shape of the queue, the AI's share of the work, the quality of every conversation and the customers to call, without exporting a thing.
Know what the AI took on
Every resolution attributed: Tore AI, AI-assisted or a person, with the hours it saved.
guessing whether the AI was worth itEvery conversation reviewed
Quality scored against your own standards, for every conversation, not a sample.
reviewing 2% of tickets by handCatch customers drifting
A risk score from real support signals flags an account while you can still help.
hearing about it in the cancellation emailWho did the work, how well,and who might leave.
Resolution attribution
Know exactly what the AI took off your team.
Every closed conversation is attributed: resolved by Tore AI, resolved by a person using Tore's draft, handled by a person start to finish, or answered by your help center before anyone wrote in. Watch the AI's share week by week, and the agent hours it saved.
- Attributed at close, for every conversation
- Self-serve counted when a help center visit ends without a ticket
- Hours saved, from your own team's average handle time
insights · resolutions
Example dataWho resolved it, by weekNorthwind · 6 weeks
Sep 21Sep 28Oct 5Oct 12Oct 19Oct 26
AI resolvedAI-assistedPerson
Agent hours saved this week38h
Conversation quality
Every conversation scored against your standards.
Write down what a good answer looks like, and every conversation is scored against it, whether Tore AI or a person answered. Disagree with a score and say why; the scoring learns from the correction.
- Your own standards: accuracy, tone, policy, resolution
- Every conversation, AI and human, not a sample
- Correct a score in plain words and it remembers next time
quality · conversation C-2291
Example data
Alex answered HanaNorthwind Mobile · sync stuck on pending
82Accurate answer95
Followed your support policy100
Tone: warm and plain55
Resolved first time80
"The tone was fine. Hana asked for short steps, so short was right." · Lena, team lead
Updated to 90. I'll treat short answers as warm when the customer asks for them.
Custom reports
Your own question, your own chart.
Pick a measure, group it, filter it and choose a chart. Save it next to the built-in reports, or export the data behind any view in one click.
- Measures for volume, times, satisfaction, AI share and quality
- Group by product, channel, tag, team or agent
- Export any report, built-in or your own
reports · new
Example dataMeasureAI-resolved share Group bytag Filterproduct: Northwind Weblast 30 days Chartbars
Save to InsightsExport CSV
Churn risk and the save offer
Spot the drift. Make one good offer.
Each customer gets a risk score from ten support signals, like rising escalations, sentiment drift and churn-intent language. You set how much each one counts. When someone does reach the cancel button, one save offer can show them what you fixed for them, with the offer you choose.
- Ten weighted signals, four risk bands, tuned to your business
- Group customers into cohorts to compare like with like
- One save offer at cancel, set up in your billing system
customers · Beacon Data
Example dataBeacon Data78high
escalation rate up30
sentiment drift25
"thinking of switching"25
reopened conversations20
In Beacon Data's cancel flow
Before you go, Priya
The export bug you reported was fixed on Oct 21, and all 4,812 rows now download. If you'd like to stay, here's a month on us.
Stay and take the monthCancel anyway
Tune it
Score quality against your standardsRisk signal weights escalations 30 · sentiment 25Save offer at cancel one offer, on
4ways a conversation resolves: AI, AI-assisted, person, self-serve
10churn-risk signals, each with a weight you can change
4risk bands: low, medium, high and critical
1click to export the data behind any report
Keep exploring.
Know what's happening,and who made it happen.
Connect your inbox and Insights fills in from the first conversation. No report to build first.