%20(1).webp)
.webp)
What has not kept pace is what becomes of a spoken announcement once the agent has finished saying it. An automated boarding call is generated from flight data, so it exists as text before anyone hears it, and that text is why the same message can also appear on a screen, be translated, or be looked up a month later. A live page starts as sound. Unless something turns that sound into text, sound is all it stays, and closing that gap is most of what we built the speech technology in Airport CX to do.
What Automation Already Solved
Most of the day runs without anyone needing to speak at all, and that is time given back rather than work taken away. Boarding calls, gate changes and delays are all sitting in the flight data before anybody thinks to announce them, so the system generates the wording itself and reads it aloud in every language the airport has chosen to cover. Change a gate and the announcement rewrites and replays without being asked. Airline names and destinations come out the way the airport settled on during setup, rather than however a text-to-speech engine would guess at them. For example, Vancouver International runs close to ten thousand automated announcements a day across eight languages, and nobody records a thing.
Each of those announcements begins as text, which is why it reaches the flight information displays at the same moment it reaches the speakers. A passenger who cannot pick the audio out of the concourse reads the wording instead. A passenger who does not speak the local language hears the same message in a language they do. For the routine day, which is to say most of the day, that is as far as the problem goes.
What a Live Page Loses the Moment It Is Spoken
Live paging exists because no set of templates covers everything, and because some messages are faster and better said than typed. Whatever the reason on the day, a gate agent picks up the handset and speaks, which is exactly what should happen and exactly what an automated system was never going to do on their behalf.
What the agent cannot do, however willing they are, is make a spoken announcement behave the way a generated one does. Set the two side by side on four counts:
- Languages. The automated announcement goes out in every language the airport configured. The live page goes out in whichever one the agent speaks.
- On the displays. The automated announcement reaches the screens from the same text that produced the audio. The live page has no text, so there is nothing to put on a screen.
- Reaching a passenger who cannot hear it. The automated announcement is readable. The live page is not.
- Afterwards. The automated announcement is logged as text, so it can be searched. The live page is logged as audio, so finding one particular call means listening for it.
All four are handled for the automated announcements and left open for the live ones, which is the wrong way round, since a live page is usually made under pressure and could matter more. This is not a matter of less care going into the live page. The live page is simply handed nothing to work with.
LIVE CALL
Try any of the three
Pick a call to place it.
SPEECH-TO-TEXT
VISUAL PAGING
Show the original transcript on the screen
Review the text, edit if needed.
AI TRANSLATION
Review the text, edit if needed.
DISPLAY & BROADCAST
VISUAL PAGING
TTS ANNOUNCEMENT
Seen and heard together — one message on every screen and speaker at the gate, in every language you serve.
PAGING ARCHIVE
Saved to the queue
Every call (original audio, visual paging, and TTS) is kept as one event.
Putting Live Announcements on Screen with Speech-to-Text
All four gaps come down to one thing: a spoken announcement is never written down. Speech-to-text writes it down. Transcription runs while the call is being made, so the agent picks up the handset and speaks much as always, and the words reach the screens at the gate as each sentence lands. That is visual paging, and it puts the live announcement in the same category as an automated one. Passengers can hear it or read it, whichever works for them.
Visual paging is also where accessibility obligations are either met or quietly missed. Canadian transport regulations name the format outright, requiring that departure, gate and safety announcements inside a terminal be made in both audio and visual form, with no exception carved out for live pages. US rules leave the method to the airport, asking instead that communication be as effective for a passenger with a disability as it is for everyone else. Neither regulation is satisfied by any single feature, but both become far easier to answer for once live pages appear on a screen rather than existing only in the air. Our airport accessibility page covers how Airport CX approaches these requirements, and the compliance page sets out the standards and certifications behind our deployments.
Translating a Live Call Into the Languages at the Gate
A screen solves the problem for anyone who could not hear the announcement. It does nothing for a passenger who heard every word and understood none of them.
That passenger has been well served all morning, because the automated announcements were reaching them in their own language the whole time. Then an agent picks up the handset and the language coverage disappears, at the one moment it was needed most.
Translation restores it, and the hard part is already done, because the page exists as text. The agent chooses which languages the call should go out in, and the translation follows within seconds, appearing on the displays, read aloud by text-to-speech, or both. A call made in English reaches the gate in French and German shortly afterwards. The judgement behind it still belongs to the person who could see the aircraft and read the room. Our multilingual announcements page goes into how language coverage works across automated and live announcements.
"The text-to-speech implementation has been a game changer for us. Before, we were chasing down different paging processes as well as video processes and trying to get them to align. With our text to speech via Airport CX, we see multilingual and ADA capabilities. It is just out of the box. It is fantastic." - Matthew Chandler, Cloud and Systems Engineer, Tucson Airport Authority
Keeping a Searchable Record of Every Announcement
Live pages do get archived. The difficulty is what they are archived as, because a recording can only be found by someone willing to listen through recordings. When a passenger writes in weeks later insisting no announcement was ever made, that is a great deal of listening for the sake of one sentence.
Transcription changes the nature of the record. A live page that was written down as it was spoken gets logged the way an automated announcement does, with its wording and the time it played, in one searchable history of everything the terminal said that day. That matters for three separate reasons:
- Answering a complaint. Search by passenger name, flight, gate or zone, and the exact wording and timestamp come back in about the time it takes to open the file.
- Evidencing a compliance review. Auditors tend to want records rather than assurances, and a searchable log of what was said and when is a record in a way that a description of your procedures is not.
- Understanding an irregular operation. Working out how a difficult afternoon actually unfolded is a different task altogether when the announcements are a list you can read through rather than hours of audio to sit through.
From the Weakest Announcement to the Strongest
Everything here depends on one step. Writing an announcement down as it is being spoken turns a live page into the same kind of thing an automated one already was, and from there the words can appear on the screens, be translated, and be found again next month.
This only works if the pieces connect. Buy transcription from one vendor, translation from another and your announcement logging from a third, and every handoff between them is a place where a live call can quietly fall out of the process. Airport CX speech technology runs transcription, translation, display and archive as one path, so a single call from a gate agent is heard, read, translated and logged in one pass.
Live pages are the weak link only for as long as they stay audio. Written down, they become the strongest announcement the terminal makes. An automated announcement knows what the flight data knows. A live page knows that and everything the person at the gate could see, and once it exists as text it reaches everyone the automated announcement reached. No airport should have to choose between a system that covers the whole terminal and a person who understands what is happening in it.
To see it working against your own flight data, request a demo.