NewKasper Voice 2.0 — native Iraqi voice

Every call answered,
in their own dialect.

Kasper answers your calls, holds the conversation, and files the ticket — in your customer’s own dialect, at the speed of a person.

Book a demoHear a real call
First-call resolution
89%
Average latency
420ms
Calls routed / mo
2.4M

Kasper agent · Baghdad

Inbound · +964 790 ••• 1824

Live
  1. Callerاكو موعد فاضي هاليومين؟
  2. Kasperإي، عدنا موعد الخميس الساعة عشرة.
  3. Callerزين، سجّلني بيه.
Stream · Kasper AI02:14
Live demo

Talk to it yourself.

No sign-up, no form. Press the button, ask what a customer would ask, and hear the answer in the dialect they would hear.

Book a demo

What is happening

The state of the call and how much of the session is left. Kasper writes the full transcript into your workspace — the demo does not keep one.

Things people ask it

  • What are your opening hours?
  • اكو موعد فاضي هاليومين؟
  • I need to reschedule my appointment.
  • شكد سعر التوصيل للبصرة؟

Customers

Trusted by operations teams across the region
  • Al-Rafidain
  • Kurdistan Net
  • Basra Post
  • TigrisLabs
  • Zaytoun Bank
  • BaghdadCare
  • Al-Warkaa
  • Shatt Logistics
The difference

The calls you are missing are the ones that were ready to buy.

Most businesses do not lose customers to a competitor. They lose them to a ringing phone.

Without Kasper

  • Calls ring out after hours, and the ones that do connect wait behind a menu.
  • Whoever answers writes the same details into a form they have already been told.
  • Nobody can say what was promised on the phone last Tuesday.
  • Arabic-speaking customers get an English line, or a callback that does not come.

With Kasper

  • Every call is answered on the first ring, at three in the morning as readily as at noon.
  • The ticket is written while the caller is still talking, with the fields already filled.
  • Every conversation is transcribed, summarised and searchable from the day it happened.
  • They are answered in their own dialect, by something that sounds like it is from here.
How it works

From first ring to finished ticket.

One flow, inside one tenant. Nothing is bolted on afterwards and nothing is copied by hand.

  1. Native audio stream

    They call, and it answers.

    Your number, WhatsApp or your own site. The caller hears a voice on the first ring, in the dialect they speak, with no menu to get through first.

  2. Kasper voice engine

    It knows your business.

    Your hours, your prices, your bookings and yesterday’s conversation are in front of it before the caller finishes their sentence — not fetched if it thinks to look.

  3. Structured outcome

    The work is already done.

    Before the call ends there is a transcript, a summary and a ticket with the fields filled in — routed, prioritised and waiting for whoever owns it.

Platform

Every capability. One tenant model.

Voice, knowledge, tickets and governance are designed together, so nothing has to be reconciled afterwards.

  • It speaks, rather than transcribes

    Audio goes straight to the voice engine. Nothing is turned into text and back again, which is where the pause in every other system comes from.

    First replyms
    Kasper420
    Speech-to-text stack1400
    Touch-tone menu2100

  • It sounds like it is from here

    Baghdadi, southern and Mosuli — warmth, laughter and a real cadence, not a newsreader reading your prices aloud.

    88%

    Baghdadi accuracy

    • بغدادي
    • جنوبي
    • موصلي

  • It replies before they wonder

    Under half a second to first word. Long enough to think, short enough that nobody asks whether the line is still there.

    30-day median↓ 24%

  • It knows what you know

    Hours, prices, bookings and documents are put in front of it before the turn, so it cannot answer from a guess.

    Connected sourcesKnowledge
    Business hours12 cards
    Customer records4.2k rows
    Product FAQ88 cards

  • Your data stays yours

    Every business is its own tenant, isolated at the database row. Access, rollout and policy sit above what runs.

    Tenant isolation18 / 24 active

  • The ticket writes itself

    Conversations become tickets with the fields already filled and the customer already attached. Nothing is typed twice.

    Drafted today172
Channels

One agent. Every way they reach you.

Voice, WhatsApp, chat and email land in the same inbox, with the same memory and the same tickets behind them.

  • Phone

    Your existing number, answered on the first ring, day or night.

  • WhatsApp call

    Not yet

    The same agent, on the app most of your customers already use.

  • WhatsApp message

    Written replies that keep the thread, the history and the ticket together.

  • Web chat

    A chat on your own site that knows what the caller asked yesterday.

  • Web voice

    Talking, in the browser, with no app to install and nothing to sign up for.

  • Email

    Answered, filed and escalated by the same agent that takes the calls.

By the numbers

420ms
Median first-token reply latency, end to end.
89%
First-call resolution on governed tenants.
2.4M
Voice minutes served per month across tenants.
6d
Average rollout time from setup to live.
Pricing

Grows with what you answer.

Start with one number and one team. Add channels, operators and workspaces when you need them.

  • Starter

    For a single team launching their first AI voice workspace.

    $290/month

    • 1 governed workspace
    • 3,000 AI voice minutes
    • Iraqi Arabic persona

  • Scale

    Most popular

    For operations teams running real customer traffic.

    $890/month

    • 3 workspaces, unlimited operators
    • 15,000 AI voice minutes
    • Knowledge and routing studio

  • Enterprise

    Multi-workspace, multi-region, governed rollout.

    Custom

    • Unlimited workspaces and operators
    • Dedicated voice capacity
    • SSO, audit log, compliance
Compare plans in full
Questions

What people ask before they buy.

Stop losing the calls you never hear.

Fifteen minutes on a call with us, and you will hear it answer one of yours.

Book a demoHear a real call