Use case / Contact centre & BPO

Answer every tier-1 call with an AI agent

AI voice agents answer the queue at volume, deflect the repetitive calls end to end, and escalate the hard ones to your human teams — with containment, sentiment, latency, and QA rolling up per line. Every call answered, scored, and measured.

Voice · Chat · Containment · Escalation · QA · CSAT · AHT

app.mirakash.com/qa
QA /Review queue
N
ScorecardsEscalationsIssues
Escalations
7
Failing
2
Avg score
4.6
Resolved
92%
#ns41
Refund escalated
+1 646 · billing
AMescalated
#ns38
Outage · West region
+1 855 · tech
PKescalated
#ns34
Tier-2 handoff
+1 628 · disputes
JDopen
#ns2f
SLA breach risk
+1 312 · warranty
AMreview
#ns27
Repeat caller
+1 917 · payment
PKopen
How the agents work

Answer, deflect, escalate, and score

A Mirakash agent answers the queue, resolves the calls it can, and hands off the ones it shouldn't — every conversation transcribed, scored, and rolled into the numbers your floor runs on.

Tier-1 voice agents

AI agents answer the queue, verify the caller, and resolve FAQs and routine servicing end to end — sub-second, at any call volume.

Deflection at volume

Repetitive, high-frequency queries — balances, hours, order status, resets — handled in full, so the queue only holds calls that need a person.

Consult-and-resume escalation

When a call needs judgement, the agent side-calls a human or another agent, gets the answer, and returns to the caller — no cold transfer, no repeat.

QA on every call

An LLM-judge scorecard grades each conversation against your rubric; a review queue surfaces the calls worth a human listen.

Vocabulary

Speak the language of the floor

Mirakash ships a small set of building blocks. Configure them to how a contact centre actually runs — calls, call flows, escalations, containment — with no forked build.

Generic in, floor terms out
AgentTier-1 voice bot · IVR replacementThe AI that answers your inbound queue.
ConversationCall · Chat · Handled interactionOne end-to-end interaction, transcript and all.
WorkflowCall flow · Routing treeVerify → resolve → escalate, as a node graph.
EscalationWarm handoff to a human teamConsult-and-resume when a call needs judgement.
AnalyticsContainment · AHT · CSAT · latencyThe CX numbers you hold every queue to.
Routing

From the line down to the right team

Model your operation as a tree: the inbound line at the top, AI agents fielding tier-1 beneath it, and your human teams at the leaves for the calls that need them. Volume and containment roll up the branches, and every team sees only its own queues.

  • Line → AI agent tier → human team → skill, as deep as you need
  • Call volume and containment roll up each branch
  • Team-level scoping enforced server-side, not in the UI
See routing & escalation
routing / call-tree
Call routing tree4 levels · containment roll-up
3
AI agents
6
Human teams
6,400
Calls/day
70.4%
Avg contained
Support Line · +1 415 555 0142Inbound line6,400 calls/day
Mirakash Tier-1 AgentAI agent · voice72% contained
Billing teamHuman queue14 agents
Payments podSkill queue6 agents
Refunds podSkill queue5 agents
Disputes podSkill queue3 agents
Tech support teamHuman queue9 agents
After-hours AgentAI agent · 24/768% contained
On-call teamHuman queue4 agents
Voicemail-to-ticketAsync queueauto
Overflow AgentAI agent · spike61% contained
Nairobi BPO teamHuman queue22 agents
Calls and containment roll up the tree — each team sees only its own queues.
Cost & usage

Cost accounted per call, to the cent

See exactly what every call costs: the STT, LLM, TTS, and telephony-transport minutes itemized per conversation, then rolled up per queue and per day. Every minute accrues to an auditable ledger you can reconcile against your provider bills.

  • Itemized STT, LLM, TTS, and transport cost per call
  • Rolled up per queue, per agent, and per day
  • Cost-per-resolution set against human cost, side by side
  • Append-only usage ledger — reconcilable to provider bills
Pricing & usage models
app.mirakash.com/numbers
Numbers /Phone numbers
N
NumbersProvisioningUsage
Numbers
42
Provisioning
3
Minutes · MTD
128k
GeoNumberTypeStatus
US+1 415 555 0142localActive
US+1 646 555 0188localActive
US+1 855 555 0142tollfreeActive
US+1 312 555 0170localProvisioning
US+1 210 555 0133localActive
US+1 628 555 0146localPorting
active · Twilio subaccount · BYO auth token AES-256-GCM encrypted at rest
In the product

Build, launch, score, and resolve — in one place

Build an agent in the no-code studio, launch it to your lines, hold every call to a QA scorecard, and work the escalations from one desk.

Build & launch

Take an agent from draft to live

Build the agent in a no-code studio, test it in the browser, and version every change. Move it through your own stages on a shared board — draft, browser test, staged, live — then connect a phone number or chat channel and answer the queue.

  • No-code builder, browser-testable, every version kept
  • Draft → test → staged → live, your own stages
  • Per-workspace isolation enforced in the data layer
Agent builder & versioning
app.mirakash.com/calls
Conversations /Live calls
N
BoardTableLive
Agent: Front desk ▾Filter8 live · avg 1:24
Listening3
+1 415 555 0142
order status · 0:14AI
+1 646 555 0188
refund · 0:39AI
+1 855 555 014
outage · 0:58AI
Thinking2
+1 312 555 0170
warranty claim · 1:22AI
+1 210 555 0133
payment · 0:47AI
Speaking2
+1 917 555 0121
address change · 0:31AI
+1 503 555 0199
cancellation · 0:20AI
Handoff1
+1 628 555 0146
tier-2 handoff · 3:05JD
Measure

A live dashboard for every queue

Turn the conversation stream into a live dashboard. Define the metrics that matter in a contact centre — containment, escalation rate, CSAT, average handle time, and response latency — and render them per queue, per agent, or per line, with a hard leak guard between tenants.

  • Containment, escalation rate, CSAT, AHT, and latency
  • Metric, time-series, and conversation-list widgets
  • Rendered per queue or line, isolated by construction
Custom metrics & dashboards
app.mirakash.com/analytics
Analytics /Tier-1 · containment analytics
N
DashboardMetricsWidgets
Containment
78%+4pt
Escalation rate
9.2%−1.3pt
Avg sentiment
+0.42+0.05
Cost / call
$0.019
Calls over time · 30dmetric: calls
Escalations · count
Cost · $/call
Recent calls
call.completed2m14s
escalationrefund
call.completed0:48
intent.detectedbooking
call.scored4.8
Resolve

Every escalation and flagged call on one desk

A single desk sits over the whole operation: the calls agents escalated, the conversations QA flagged, and the ones awaiting a human. Route by priority, resolve with the full transcript and context, and feed every outcome back into the scorecard.

  • Escalation inbox, review queue, assignment, and resolution
  • Consult-and-resume handoffs keep the caller in context
  • Resolved outcomes flow back into QA and analytics
Review queue & QA
app.mirakash.com/qa
QA /Review queue
N
ScorecardsEscalationsIssues
Escalations
7
Failing
2
Avg score
4.6
Resolved
92%
#ns41
Refund escalated
+1 646 · billing
AMescalated
#ns38
Outage · West region
+1 855 · tech
PKescalated
#ns34
Tier-2 handoff
+1 628 · disputes
JDopen
#ns2f
SLA breach risk
+1 312 · warranty
AMreview
#ns27
Repeat caller
+1 917 · payment
PKopen
Team console

Give every team its own console

Each team works from its own console on its own domain — the live escalation inbox, the calls it owns, QA standing, and usage — resolved only for verified members, so access never widens.

White-label portals
help.northstar.io
Northstar Contact
ChatOrdersSupportStatuspowered by Mirakash
<1s
Avg reply
72%
Contained
280
Agents
Live chat
Where's my order? It's late.2:14Customer
Out for delivery — arrives by 6pm today.2:14AI
Great, thank you.2:15Customer
Your brand · your domain · one agent across web chat, SMS & WhatsApp.
Outcomes

What a floor on Mirakash looks like

182,400
calls handled / month
72.4%
tier-1 containment
740ms
avg response latency
4.4 / 5
avg QA score
Containment ratetracked
AHT & latencytracked
QA & reviewtracked
CSAT by queuetracked

Ready to put an AI agent on every call?

Book a walkthrough and see the voice pipeline, telephony, workflows, QA, and human handoff working together — on one platform.