Maria closes deals on calls. She is sharp, present, commanding. Her last performance review noted "exec presence on calls" as a particular strength. Then she opens Slack to follow up, and something changes.
The hidden tax of writing in your second language is not vocabulary or grammar. It's the time you spend translating your confidence from speech into text. This gap is sharpest for senior knowledge workers who speak English fluently but write it carefully, checking their words three or four times before sending.
The moment it shows up
Maria has been selling into US enterprise for eight years. She speaks English every day, often more hours than her native language. On a video call, she's fluent, fast, idiomatic. She lands jokes. She challenges assumptions. She closes deals.
Then she has to write a follow-up Slack message to her team. She opens the message box. Types what she wants to say, something quick, a two-sentence update on deal status. She reads it. Deletes it. Rewrites it. Reads it again. The message looks smaller than it sounded in her head. Less direct. Less her.
This cycle repeats. She rewrites. Four minutes in, she's satisfied enough to send. Total time on a two-sentence message: thirty minutes, because she's spent fourteen cycles editing.
Later that week, her manager gives feedback. The call performance was strong. The written communication needs work, "Could be more concise and direct in Slack." Maria knows the gap isn't real. She's not less concise on calls than in Slack. Something else is happening: her voice translates differently into text.
This isn't a rare moment. For non-native English speakers in senior roles, this is a weekly tax. The pattern: speak fluently, write carefully. Sound sharp on calls, sound cautious in Slack. Spend three times longer on written communication as you would on spoken communication for the same idea.
Why this isn't a grammar problem
Most tools for non-native English speakers assume the problem is English itself. Grammarly catches syntax errors. Translation layers try to make your writing sound more "native." But these tools miss the actual friction.
Maria doesn't need help with English. She's fluent. She speaks it confidently on calls, in meetings, in presentations. What she needs is to translate that confidence into writing without losing it in the translation.
When you write from scratch, you make different choices than when you speak. You slow down. You second-guess. You write to a perceived standard of English instead of how you think. The careful version of yourself shows up on the page.
On a call, your accent is part of your presence. Your pace and confidence are audible. In Slack, you have only text. No signals of fluency. So you rewrite. You polish. You try to sound more correct, less regional. And by the time you hit send, you've spent ten minutes on a message that took two minutes to think through on a call.
The speech-to-text bridge
Whisper, OpenAI's speech-to-text model, has gotten very good. The large-v3 version hits 96.3% word accuracy on LibriSpeech, which means for most professional sentences, you get a usable transcript on the first pass. The transcript is messy, filler words, pauses, the um and uh that you don't notice in speech. But the core is there.
Then a polish step removes the filler, maybe reorders a clause, tightens the sentence. But the emotional core of what you wanted to say is already captured. You're not translating from thought to formal writing. You're cleaning up a transcription of how you actually talk.
Why this works for non-native speakers is subtle. Your accent, pace, and linguistic patterns when speaking are preserved in the transcript. When you write from scratch, you lose these patterns. You write to an internalized standard of English. The transcript captures your thinking patterns. Then you polish the grammar, not the voice.
What Recitey does differently
Recitey isn't Otter.ai or Wispr Flow, which give you a messy transcript and ask you to edit. It runs Whisper locally on your device with no cloud lag. Then it polishes the transcript in about two seconds, removes filler words, fixes obvious errors, handles minor grammar. What you get is a clean sentence that sounds like you, not one that sounds like you've spent an hour rewriting it.
No word limits. No monthly cap. No metering. It works in Slack, email, browsers, Google Docs, your terminal, anywhere on Windows where you can type. You speak. It cleans. You paste.
The difference from Grammarly is instructive. Grammarly catches your mistakes. Recitey gives you back your time. Grammarly assumes you wrote something that needs fixing. Recitey assumes you spoke something that needs polishing, a subtly different starting point, which changes everything downstream.
How the workflow actually changes
For Maria, the shift is concrete:
Before voice: Open Slack. Type a message. Reread it. Delete it. Retype it. Reread it again. Rewrite it once more. Seven to ten minutes of editing cycles, four or five iterations.
After voice: Open Slack. Speak the message for twenty seconds. Two-second polish. Paste. One minute total.
The performance review gap doesn't close because the writing is objectively better. It closes because she stops writing cautiously. She stops rewriting to sound more "correct." The message that goes into Slack carries the confidence she has on calls. And that's what shows up in how her team experiences her.
The real value isn't faster typing
The speed matters. Reclaiming thirty minutes a week of mental energy is real. But the deeper value is that your written voice starts to sound like your spoken voice. You stop being two different people, the confident one in calls, the careful one in Slack. The medium stops forcing a translation. Your presence travels into text the way it travels into calls.
That's what "exec presence" actually means. It's not eloquence or formality. It's being recognizably yourself, confident and direct, whether someone is hearing you or reading you.