Product / Console
See every conversation, and what to fix next
Monitor conversations, track your goals and see where agents need help. Review the evidence and decide what to improve next.
Insights
Goal-linked insights and recommendations driven directly by KPI runtime state.
Production release evidence
Generate closed-day reportsSampling Bias
0.3 pp
412 selected / 40,118 eligible
Control
Exact release comparison
v12 stable
Scoring Complete
100%
closed day 26 Sep
Active Goals
4
Risk Goals
1
Open Insights
6
Recommendations
3
Insight queue
Ranked by goal risk, insight pressure, and backlog
at riskFirst contact resolution
Resolve billing disputes on first contact
Attainment
64%
71.2% / 85%3 insights2 recommendations
on trackContainment
Keep connectivity faults contained
Attainment
88%
81.0% / 85%2 insights1 recommendations
activeAbandonment
Cut WhatsApp abandonment
Attainment
40%
9.8% / 6%1 insights0 recommendations
Resolve billing disputes on first contact
Evaluate Generate insights Generate recommendationsCurrent / Target
71.2% / 85%
Baseline
66.0%
412 samples
Last Evaluation
at risk
26 Sep
Delta from baseline +5.2 ptsDistance to target −13.8 ptsSample sufficient
Goal insights
tool failureopen
Dispute evaluation times out on peak evenings
38 conversations since Monday ended in a handoff after billing.evaluate_dispute exceeded its timeout between 18:00 and 21:00.
Severity 0.72Occurrences 38Rank 0.81
Recommendations
Raise the evaluate_dispute timeout to 8 s at peak
pending reviewKeeps the customer on the line instead of handing off while the billing system is slow.
36 of the 38 timeouts completed within 8 s once retried.
Projected impact +4.0 to +6.5 ptsConfidence 62%
See the queue as it happens
Every conversation lands in Console the moment it ends, with its outcome, sentiment and duration, and the live ones while they run.
Deployed agents, active conversations, resolution rate and handle time per environment, with the agent performance table beneath.
Every ended conversation, filterable by handler, channel, resolution and time period, with sentiment estimated for each.
Open any ticket and read the transcript turn by turn, with the outcome, sentiment and handle time beside it.
A live taxonomy with classifier profiles, top intents by volume and outcome, and a review queue where assignments are confirmed, corrected or dismissed.
Dashboard
Deployed agents and customer performance in Production.
6
with an active release
128
right now
71.2%
all ended conversations
4m 12s
Voice · Chat 6m 40s
Agent performance
| Agent Name | Status | Conversations | Resolution Rate | |||
|---|---|---|---|---|---|---|
| Telecom Billing & Disputes | Stable | 40,118 | 37 | 6.2 | Voice 4m 12s | 74.1% |
| Fibre Outage Triage | Canary | 8,204 | 12 | 4.8 | Chat 6m 40s | 68.5% |
| SIM Swap & Activation | Stable | 12,930 | 9 | 5.1 | Voice 3m 58s | 70.3% |
Resolution Rate — Last 7 Days
Resolution Rate Target (80%)Set the goal, measure the gap
A KPI goal names the agent, the metric, the target and the baseline window. Ruhu measures the rest in Production.
Name it, choose the agent and metric, set the target value and date, and let the baseline be measured rather than typed.
Current against target, baseline, delta, distance to target and sample sufficiency, updated as conversations close.
Goals ranked by risk, insight pressure and backlog, with each goal’s open insights and pending recommendations.
Acme Telecom · Console
KPI Goals › New goal
Create KPI Goal
Goal Name
Description
Agent
Metric
Target Value
Target Date
Baseline Window
Manual baseline
Off · measured from Production
Act on what it finds
Insights become recommendations with a projected impact and a confidence score, and a person decides.
Each carries a projected impact range and a confidence score. Mark it ready for review, then approve or reject it.
Sampling bias, the control release and scoring completeness for every closed day, so a comparison can be trusted.
The audit log exports as CSV from Console and as CSV, JSON or NDJSON from the API, so numbers can be checked outside Ruhu.
Acme Telecom · Console
KPI Goals › Goal › Recommendations
Resolve billing disputes on first contact
Generated against measurable goal gaps
Recommendations
Raise the evaluate_dispute timeout to 8 s on peak evenings
ready for reviewProjected impact +2.1 to +3.4 ptsConfidence 0.81
ApproveRejectOffer the goodwill credit before asking for the account number again
draftProjected impact +0.8 to +1.6 ptsConfidence 0.62
Ready for reviewConsole by the numbers
8
KPIs in the catalogue, from first-contact resolution to abandonment
Every turn
Traced with its reasoning, and sentiment scored from the transcript afterwards
Production
Where business KPIs are measured, per operating environment
How it works
Up and running in three steps
01
Set your goals
Create a KPI goal with a target. Console measures the baseline from real conversations.
02
Watch and review
Monitor live calls, and open any conversation for its transcript and reasoning.
03
Act on recommendations
Take a recommendation to Studio, ship it as a version and watch the goal move against its baseline.
FAQ
Common questions
Can a supervisor take over a live call?
Not yet. Live Calls lets you watch and end a call. A handoff to a person carries the full package to your help desk.
Does Console work with our help desk?
Handoffs create tickets in Zendesk today, with Freshdesk and Jira adapters, and a handoff is only marked accepted once the provider shows the ticket exists.
Are the sentiment scores real?
Scored from the transcript after the conversation. Until then a score is marked estimated, from how the conversation ended, and shown muted.
Can we get the data out?
Every conversation, trace and audit event is available through the API, and audit logs export as JSON or CSV.
Start with one workflow
Which customer request should we tackle first?
Bring a procedure and the systems it touches. We’ll walk through the agent, the setup and how you would measure the result.