Skip to content

The traveller leaves the call
with a seat, not a case number.

Rebooking a cancelled flight and re-issuing the ticket on the same call. Voice AI Agents with Real-Time Translation, on the telephony and reservation system you already run.

Book a Demo Follow this call, step by step
  1. 01ClientsAllegiant · Amadeus
  2. 02The taskRebooking, re-issue, refund by policy
  3. 03The callSix turns, one receipt
  4. 04SystemsPSS, fare rules, ticketing
  5. 05ProofWhat a rebooking build looks like

An illustrative call plays beside this headline. The agent checks the fare rules, rebooks the traveller onto the 07:10 in seat 14C, and hands the refund to a person with the full context. A receipt of what the call changed prints at the end.

Clients01 / 05

Who runs it in travel, and what they run.

The task02 / 05

A cancelled flight is three jobs in one conversation.

Rebooking, re-issue, and the refund that needs a person. One caller wants a seat in the morning, a ticket that is valid when they board it, and money back on the fare they lost, and each of those lives in a different system.

On the runtime, in one call

One call, on the runtime.

The fare rule is read before an option is offered, and the change is written to the booking before the call ends.

Three steps, one call

Answered on the first call

The booking is found and the disruption read, so the caller is answered on the call they made.

First call

Checked, held, re-issued

The fare rule is read, the seat is held and the ticket is re-issued in one pass.

Within the rule

Refund to a person, receipt sent

The refund goes to a person with the call’s context attached, and the receipt is sent.

Handed over
The call03 / 05

The call ends with a record of what it changed.

Every answer cites the rule it came from, and every write names the system it went to. The refund is the one thing this agent does not decide, so it leaves with a person and the call’s context attached.

Illustrative call, not a recording.

Customer

My flight tonight was cancelled. Can you get me on the morning one?

Rebooked 07:10 · ticket re-issued · refund case opened

Moment 1 of 6: Asks for the morning flight, Telephony

Systems04 / 05

Your PSS stays the system of record.

The runtime runs inside your perimeter and writes only through interfaces you already expose. The fare rules are read, the seat is held, the ticket is re-issued, and nothing else in the estate is touched.

Your PSS stays the system of record.The runtime runs inside your perimeter and writes only through interfaces you already expose. The fare rules are read, the seat is held, the ticket is re-issued, and nothing else in the estate is touched. The steps, in order: Your telephony, Voicing runtime, Fare rules, Reservation system, Ticketing, Messaging, A person, Call record.YourtelephonyYour own trunksVoicingruntimeIn your estateFare rulesRead: the rulesReservationsystemRead: the seatsTicketingWrite: re-issueMessagingItinerary sentA personCall recordEvery step keptIn your perimeterNothing leaves itFare rulesRead only, never writtenTicketingOne write, declaredA personRefunds stay with them

Step 1 of 6: The call arrives on your own numbers, in your own queue.

Systems

Amadeus Altea, Sabre, FlightAware and your loyalty stack, and whatever else you run.

One runtime, inside your perimeter

  • Amadeus Altea

  • Sabre

  • Navitaire

  • FlightAware

  • OAG

Your loyalty stack answers on the same call: Loyalty Prime, IBS iFly or a custom FFP, read live, with vouchers, bags, seats and fare differences written back. Voice payments run in a subprocess of their own.

System marks belong to their owners.

And it does not end here.

  • A documented API

    Every action the agent takes, as a call.

  • An MCP server

    Your tools, over the Model Context Protocol.

  • Webhooks

    Events pushed to whatever listens.

  • Any system with an interface

    Your own disruption tooling, same terms.

What it is held to.

  • Cardholder data

  • Control audit

  • Data security

  • Data residency

    Audio residency

Proof05 / 05

One travel deployment, and what it measured.

  • 65%lower operating costThomas Cook, inbound itinerary and disruption calls
  • 80%of itinerary changes doneThomas Cook, date, route, name and ancillary changes
  • 12languages handled nativelyThomas Cook, European passengers, Real-Time Translation

What changed, before and after go-live.

Seven measurements, each with its before. Every row is one deployment’s own figure on both sides of go-live, and no change is stated as a percentage because the deployment never published one.

  1. Average handle time8.1 min2.7 min
  2. First call resolution38%74%
  3. Cost per interaction$9.80$1.40
  4. CSAT score3.2 / 54.3 / 5
  5. Agent escalation rate84%21%
  6. IROPS containment rate22%71%
  7. Rebooking self-service rate18%67%

One travel deployment’s figures as published by Voicing AI; period on request.Ask us for the report

Bring us one route’s worth of disruption calls.

An agent hub joined by tubes to a database, a ticket tray, a calendar and a vault behind a small gate, with a receipt tile beside it.
Airlines

Rebooking a cancelled flight, end to end

Book a review with the architects who would run it, not a sales call. We will read your call flow, name the systems it touches, and tell you what we would not automate. Then take a week of disruption calls and count the rebookings a person had to correct, and the refunds that reached a person without their context.

Book a Demo

Request the security package