Host a town hall or all-hands

Town halls present a specific challenge for multilingual organisations: the room is large, the speaker roster rotates, and the attendees span language groups that the company’s working language may not serve equally. (See Corporate town halls for an overview of how organisations are using realtime translation in this format.) Traditional simultaneous interpretation scales poorly to the format — it requires booths, headsets, and a booking lead time that does not match the typically short planning horizon of internal communications.

Loquira fits the format because it shifts the infrastructure requirement from the room to the listener’s pocket. This guide covers the operational adjustments that make it work at town hall scale.

Coordinating with AV

Most town halls happen in rooms that already have a sound system — a microphone at the podium, house speakers, a mixing desk. Loquira’s presenter view captures audio from the same microphone that feeds the house sound. The key rule: the microphone that feeds the room must also feed the browser. Review the audio requirements to ensure the signal meets recognition thresholds.

The simplest setup: connect the podium microphone (or the mixer’s auxiliary output) to the presenter’s laptop via a USB audio interface. The presenter runs Loquira in the browser, which captures the audio from the interface. The house sound system is unaffected — the audience hears the speaker through the room speakers; remote attendees receive both the original audio and the translated stream through Loquira.

If the AV team is reluctant to route a signal to a laptop, the fallback is a dedicated presenter microphone — a lavalier worn by the speaker, feeding a second channel into the Loquira browser session. The speaker speaks into the room mic; the lavalier captures the same audio for the recognition engine. Brief the speaker to treat both microphones as active.

Pre-promoting the QR

The join URL and QR code should reach attendees before they enter the room. Three channels, in order of effectiveness:

1. Calendar invite. Embed the QR or join link in the meeting description. Attendees who open the calendar event on their phone can join immediately.
2. Email reminder. Send the QR with the pre-read materials 24 hours before the town hall. Include a one-sentence explanation: “This meeting will be translated in real time. Open this link on your phone and select your language.”
3. Slack or Teams pin. Pin the join link to the relevant channel an hour before the event. This captures late adds and attendees who cleared their calendar.

On the day of the event, display the QR on the main screen during the pre-roll and hold period. Not everyone reads email before walking into the room.

Multiple speakers

Town halls typically rotate through an executive line-up: CEO, CFO, department leads, Q&A with the floor. Each speaker change creates a potential break in the translation pipeline.

The simplest solution is a single house microphone — one podium mic that whichever speaker is active uses. For guidance on microphone types and placement, see Choose the right microphone. The recognition engine adapts to a new voice within a few seconds and does not require a session restart. Brief each speaker to begin with a complete sentence at the microphone before launching into their content. The engine uses those first few words to tune its acoustic model.

If speakers rotate away from the podium — walking the room with a handheld microphone — the handheld feeds the session. The same adaptation rule applies: a few seconds of clean audio at the start of each segment is sufficient.

The presenter view operator (a designated colleague, not the current speaker) monitors the live transcript during transitions. If the transcript drifts during a hand-off, the operator alerts the speaker to return to the microphone and repeat the last point.

Post-event transcript distribution

The multilingual transcript is the most valuable output of a translated town hall. For the full cleanup and attribution workflow, see Curate transcripts after the event. It serves employees who could not attend live, those who followed in a different language and want to check the original wording, and managers who need to cascade specific messages to their teams.

Steps for distribution:

1. End the session deliberately to trigger the export build.
2. Download the ZIP archive containing all language transcripts.
3. Upload the ZIP to the company intranet or knowledge base, or attach it to the meeting follow-up email.

The transcript is a record of what the speech engine recognised. For formal communications — policy announcements, compliance updates — compare the transcript against the speaker’s prepared remarks before distribution. The engine may have missed or substituted words during periods of cross-talk or audience applause.

Measuring engagement across language cohorts

After the event, the session dashboard shows per-listener connection timestamps and language selections. This data answers a question most internal communications teams cannot currently answer: which segments of our workforce are not fully served by the company’s working language, and how long did each listener stay?

If a town hall with 500 attendees shows 80 listeners in Spanish, 45 in French, and 30 in Japanese, that is actionable data. It tells the communications team that those cohorts exist and that they chose to follow in their mother tongue when given the option. Over successive town halls, the trendline reveals whether adoption is growing and which languages to prioritise for future sessions.