HR: how many candidate voice notes do you play per role?
Count last week's. The first one lands ten minutes after the vacancy goes out: three minutes and twenty seconds of somebody walking through their previous job, with traffic in the background. Four more follow. The hiring manager sends a voice note asking you to move the spreadsheet requirement from nice to have to essential. On Friday, a candidate who did not make the cut sends a voice note asking what was missing.
No single one of those is a problem. Together they are a queue — and the queue shows up in no report.
Voice is not the exception in screening. It is the default
Anyone who recruits in a WhatsApp-first market worked this out without a survey. The scale still helps: in March 2022, WhatsApp said its users send an average of 7 billion voice messages a day. Speaking is faster than typing while standing in a queue — which is exactly when many candidates reply.
Brazil, where voice notes at work went mainstream early, gives a useful picture of how lopsided it gets. Opinion Box's Pesquisa WhatsApp no Brasil, run in June 2025 with 1,126 users and a 2.9 percentage-point margin of error, found text messaging still first at 88% — with voice notes right behind at 80%, ahead of voice calls (72%) and video calls (63%).
On the employer side the channel is no longer informal either. Cetic.br/NIC.br's TIC Empresas 2024 survey found 49% of Brazilian companies used WhatsApp, Skype or Facebook chat messaging that year, down from 55% in 2023. Half the employer market talks there — hiring included.
The arithmetic nobody times
Here is an illustrative calculation, not survey data — just the maths of an ordinary week. One administrative assistant role, five shortlisted candidates. Each sends a two-to-three-minute voice note about their experience. That is 10 to 15 minutes of pure listening. Add the manager's voice note changing the brief, and two follow-up questions that arrive later, and the clock passes 20 minutes before a single interview happens.
Listening time is the small part. What eats the day is the flow around it: open the chat, download the file to hear it properly, put the headphones on, write the previous employer's name down by hand, scrub back twenty seconds because you missed it, drop everything when the phone rings, start again from zero. Now multiply by six open roles.
And that cost lands on the tightest part of the process. In SHRM's 2025 benchmarking reports, based on data collected between January and March that year, screening and interviewing each average 8 to 9 days inside the hiring process. Every half hour handed back to screening is half an hour that returns to talking with people.
What exists today — and where each route stops
Before installing anything, it is worth knowing what WhatsApp itself and plain common sense already solve. And where each route hits its ceiling.
| Route | What it solves | Where it stops |
|---|---|---|
| Playing at 1.5x or 2x | Cuts about a third off listening time, nothing to install | A noisy recording or a strong accent turns to mush — and you still have no text to paste into the ATS |
| WhatsApp voice message transcripts | Announced by WhatsApp in November 2024; the transcript is generated on your device | The announcement mentions a few select languages at launch, and the feature lives on the phone — not in the WhatsApp Web panel where screening actually happens |
| Asking candidates to write instead | Standardises answers and makes comparison easy | It lowers reply rates exactly among the people with least time and no keyboard; part of the queue never comes back |
| A recruiting chatbot on the company number | Organises whoever enters through the bot's funnel, with fixed questions | The voice note that lands in your direct chat with a candidate or a manager is still yours to handle |
| Typing it out yourself | Comes out exactly the way you want it | Three minutes of audio costs ten to fifteen minutes of typing; nobody sustains that across five roles |
None of these routes is wrong. They all stop at the same place: the voice note that arrived today, on the screen you work on, still has no text.
How this works in ZapVox, step by step
ZapVox is an extension for WhatsApp Web on your computer. It reads the voice note that arrives in the conversation and puts the text right there, under the message. For screening, the path looks like this:
- Turn on automatic transcription. New audio arrives already transcribed, no click needed. When you sit down after a meeting, all five voice notes for the role are text under each message.
- Ask for the bullet-point summary. A three-minute voice note becomes a few lines with what matters: where they worked, for how long, what they do now. Enough to decide who gets an interview.
- Use per-chat settings. In each conversation's gear icon you set language, translation and summary for that chat only. The internal group stays out of it; the chat with an overseas candidate comes through translated.
- Search inside voice messages. Need to find who mentioned a specific system? Search runs over the transcripts stored in your browser and shows which conversations the word appeared in, grouped by person, opening the chat at the right spot.
- Export what you need on record. A conversation's transcripts come out as a .txt file, ready to paste into the ATS or the role's spreadsheet.
- Schedule the follow-up. The message to candidates who did not make it can be written at 10pm and go out in business hours. Before sending, ZapVox double-checks the recipient — which matters when seven chats are open.
One detail changes screening more than it sounds: with two candidates' text side by side, you compare by reading. It takes about as long as playing one of them.
Where this stops
The honest part, because it exists.
- It is on the computer. ZapVox runs on WhatsApp Web in Chrome, Edge, Brave and Firefox. If your screening happens entirely on the phone, on the bus, it will not follow you there.
- The browser has to be open. Automatic transcription happens while the tab is there. Computer off, nothing is transcribed — the audio simply waits in the chat.
- There is a daily allowance. The Free plan gives 10 transcriptions a day and up to 15 minutes of audio a day. During the 7-day trial, 15 operations a day and up to 30 minutes of audio. A peak week with four open roles can hit that ceiling.
- It does not walk into your ATS by itself. The text lives in WhatsApp Web; getting it into the system is a copy and paste, or the .txt export.
- Bad audio makes bad text. Street noise, two people talking at once, a choppy recording: the text comes out with gaps. On a decisive passage, confirm by listening.
- A transcript is a working tool. It is there to read, compare and keep a record of what was said. It is not evidence, and it replaces nothing your legal team requires in writing and signed.
- It does not transcribe interviews held on a call. What it reads is voice messages arriving in the conversation.
If you want to test it on a real role
The honest way to decide is to run it on an actual vacancy, with the queue of voice notes you already have. Every sign-up starts with 7 days of trial, features open and no credit card. When it ends, the account returns to the Free plan on its own — nothing is charged. Plans and their limits are on the plans page.
Installing ZapVox on Chrome takes less time than playing the first voice note in the queue.
In two lines
- The problem is not the candidate's voice note — it is the flow around it: download, play, write down, replay.
- With the audio as text, screening becomes reading: you compare, search by word and keep a record without replaying anything.