You're excellent on client calls. Sharp, present, authoritative. Then you draft a Slack message to your team and you re-read it four times before sending. Something about the written version sounds smaller, more careful, less like you. So you rewrite it. Again. This isn't a productivity glitch. It's a medium problem that affects how non-native English speakers experience their own competence.
The Hidden Tax
This isn't a vocabulary problem. (You already speak English fluently.) It's not a grammar problem. (Most of what you write is technically correct.) The real cost is time, and it's significant enough to shape how your manager perceives your contribution.
Maria, a senior account executive at a SaaS company in Stockholm, calculated it recently: the average async message to her US team took her 8 to 10 minutes to write, then another 10 to 15 minutes of rereading and rewrites before she felt it was ready to send. On a day with 12 such messages, that's roughly 4 to 5 hours lost to async English drafting. Not on calls. Not in presentations. On text that moves at human speed.
The asymmetry was visible in her performance reviews. "Exec presence on calls" was listed as a strength. She had closed deals. She had influence. "Could be more concise in writing" appeared as a gap. Maria knew the gap wasn't real. It was just the medium stripping away the presence that survived her accent.
Non-native English speakers in sales, product management, and engineering face a specific kind of friction that native speakers don't. On a call, you are yourself. Your cadence, your thinking, your authority comes through. The accent is just part of how you think out loud. In Slack, you are alone with words. So you rewrite them. You make them safer. Smaller. More formally correct. The written version is not more accurate. It's just less like you.
This cost compounds across a team. If a senior person is spending 4 to 5 hours a day on async English drafting, that is thousands of hours annually on a single friction point. And the real damage is invisible: the ideas that don't get written because they would take too long. The quick clarification that becomes a 20-minute project. The unasked question.
The Rewrite Trap
Most non-native English speakers reach for Grammarly when they notice the friction. It catches the comma you missed, flags passive constructions, corrects typos. It is genuinely useful for correctness.
It does nothing to shorten the 20-minute drafting cycle.
Why? Because the friction isn't grammar. It's identity protection. On a call, your accent is just part of how you think. In Slack, the words are all that exists. So you rewrite them to sound more formal, less marked, safer. Grammarly polishes the careful version. It does not return the 20 minutes to your calendar.
The trap deepens because Grammarly's feedback feels productive. You fix the issues it flags. You feel more correct. Your confidence goes up slightly. But underneath, you're still rewriting the same message three or four times before you send it. The tool has optimized for a problem that was never the real constraint.
Translation tools (DeepL, Google Translate) make this worse. They erase the voice that survives your accent. You speak a thought naturally, paste it into a translator, and it comes back in flattened, corporate English that sounds like no human would ever write it. Now you have to rewrite it anyway, but you've lost the starting voice.
The Voice-to-Text Shortcut That Doesn't Work
At some point, most people try voice dictation. Microsoft Voice Typing. Apple Dictation. Otter.ai. You speak the message naturally and the tool transcribes it. For a moment, this feels right. Your actual words are there.
Then you read it back.
The draft is rough. Missing punctuation. Run-on sentences. False starts. Fragments where you were thinking out loud. "Um" and "uh" scattered throughout. So you rewrite it anyway, and now you've added a transcription step on top of the rewriting step. You haven't saved time. You've just fragmented the process into three parts: speak, transcribe, rewrite.
Some people do try to skip the rewrite step. They send the rough transcription as-is. It reads like someone was thinking out loud, which is technically what happened. But it lacks the polish that signals you took time to compose the message. So you rewrite it.
The tools that exist assume you are a beginner who needs help with basic dictation. They're not built for someone who is already fluent but is writing in a non-native language and wants to skip the rewrite step entirely.
Why Transcription Alone Isn't Enough
Transcription is just the first step. What separates a rough voice memo from a readable message is the cleanup: proper punctuation, capitalization, sentence structure, removal of filler words and false starts.
Most voice dictation tools hand you the raw transcript and let you figure out the cleanup yourself. That's where the rewrite happens. You read through, mentally translate the fragmented voice memo into readable prose, and type the corrected version.
It's still rewriting.
The fundamental issue is that tools have treated voice dictation as a transcription problem, not a voice-to-writing problem. They've optimized for accuracy (words in, words out) but not for the transformation that actually matters: taking your spoken clarity and converting it into written clarity without losing your voice.
The Recitey Approach
Recitey uses Whisper, an open-source speech-to-text model that achieves 96%+ accuracy on English speech, and polishes the rough draft into a clean, readable sentence in under 2 seconds. No word limit. No metering. No monthly cap. It runs locally on your device.
The difference is fundamental. You speak at your actual speed. Your thoughts come out in your actual vocabulary. The tool makes it readable without making it cautious.
Maria tried it on her next client follow-up. She needed to explain a proposal timeline to the account team. Instead of drafting it in Slack (which would have taken 8 to 10 minutes), she recorded a 30-second voice memo explaining the context in her actual voice. Whisper transcribed it with 96% accuracy. The cleanup pass took 2 seconds. She sent it. It sounded like her.
Compare that to her usual process: 8 to 10 minutes of drafting, then 15 minutes of rereading and rewrites before the message felt safe enough to send. Now it was 30 seconds of speaking plus 2 seconds of polish. The time savings is real, but the more important change is that the voice doesn't get transcribed and then rewritten into a more formal version. It stays sharp.
What Changes After
When async English stops costing you 20 minutes per message, the math of your day changes. Not because you are suddenly faster at typing. But because you are not fighting the medium anymore. Your written voice carries the same presence your spoken voice does. The identity gap closes. The performance review gap closes too.