Maria's performance review hit the mark and missed it entirely. "Strong executive presence," it said. "Consider more directness in async communication." She's been with the company seven years. On sales calls with US enterprise buyers, she's sharp, fast, commanding. In Slack, she's not.
That's not a skill gap. That's a medium problem wearing a skill-gap mask.
The 30-Minute Tax
When you're not writing in your native language, something changes. You can feel it. On a call, Maria thinks in Swedish, then speaks English without translation. The accent survives. Her actual voice survives. But in Slack? She writes a sentence, reads it back, deletes it. Writes another. Checks the phrasing three times. Reads it once more. Then sends.
That loop, the re-read, the hesitation, the rewrite, costs time. Not because she's slow at English. Because she's protecting herself. Written English is where bilingual professionals feel exposed in a way speech isn't. The voice gets smaller. The certainty gets hedged. The person who closes deals sounds tentative in async.
Last week, Maria spent 47 minutes rewriting a six-message Slack thread about pipeline delays. Six messages. The thoughts were the same ones she'd already explained on a call three days earlier, where she sounded authoritative and clear. The medium changed the voice.
For senior knowledge workers in DACH and the Nordics, this isn't unique. It's systematic. And it's never measured.
Why Grammarly Doesn't Solve This
The instinct is obvious: use a writing tool. Grammarly catches grammar, suggests phrasing, makes your English sound more native. But Grammarly's designed for people who're learning English or native speakers trying to sound more polished. It optimizes for formal correctness, not for recovering your actual voice.
For bilingual professionals like Maria, Grammarly can actually deepen the problem. It catches real errors. Fair. But it also "fixes" choices that were intentional, phrasing that sounded natural when you wrote it. And it does this without understanding that you already know English well enough to have strong opinions about your sentences. You don't need a tool that teaches you English. You need a tool that listens to how you actually speak and gets it into Slack without the re-read loop.
Voice Tools Exist, But They're Built for Transcription, Not for Slack
There are voice-to-text solutions: Wispr Flow, Otter.ai, Dragon NaturallySpeaking. They all transcribe speech accurately. Whisper-large-v3, the model most of these rely on, achieves 96.3% word accuracy on standard test sets. That's solid.
But here's the gap: those tools transcribe meetings, voice memos, interviews. They're built around the assumption that your voice is the primary input and Slack is an afterthought. You record or dictate into their app, get a transcript, then copy-paste it into Slack. Two steps. Two context switches. And by the time it lands in Slack, it's already become "a transcript," not "your voice as a message."
Worse, if you're working in Slack on a Windows machine, you end up using Apple Dictation or Microsoft Voice Typing as a workaround, tools that have existed since 2012 and haven't evolved for knowledge workers who need precision. They catch every background noise, every hesitation filler, every "um." You still end up rewriting.
The tools solve transcription. They don't solve the identity problem.
Speaking Into Slack, Not Writing Into It
The wedge is simple: bypass the written-English hesitation entirely. Speak the message the way you'd say it on a call. Let the tool transcribe it cleanly and land it directly in your Slack draft, in context, without the extraction step.
What changes is immediate. Maria speaks a follow-up message the way she'd say it to a client face-to-face. Within seconds, it's transcribed and sitting in her Slack drafts. She's not spending 30 minutes re-reading and rewording anymore. She's spending two or three minutes reviewing what got transcribed, making sure nothing important got tangled, and sending it.
The message sounds like Maria. Not the small, careful version. The actual one.
The technical piece matters here. Recitey runs Whisper-large-v3 locally on your Windows device via the system clipboard, with zero variable cost, no word limits, no metering per message, no subscription escalation as you use it more. You dictate into the app, it processes on your machine, and the transcript lands directly in Slack as a draft. No cloud roundtrip. No extraction step. No data sent anywhere. It's fast because it runs on your GPU. It's private because it never leaves your device.
The Real Trade-Off
This isn't a shortcut that lets you skip proofreading. You still read it before sending. You still catch the moments where the transcription misheard you or where your thought got tangled. But the frame shifts: instead of "write carefully to sound smart," it becomes "speak naturally, then proof it." That's a smaller cognitive load. Especially after six calls and three customer meetings in English.
After You Adopt It: What Actually Changes
Here's what Maria noticed in her first two weeks.
The review step is faster. You're not fixing tone or phrasing or tense, those come out right when you speak them. You're mostly just confirming the transcript heard you correctly. A message that used to take 8 to 12 minutes of rewrites now takes 2 to 3 minutes to review. That's roughly 12 to 18 minutes saved per week for someone writing four or five async messages daily.
You write more async. That sounds obvious, but it's not. When async writing feels cognitively expensive, you default to Slack threads and back-and-forth messages instead of the one clear, complete message that would actually move the conversation forward. When the friction drops, the channel behavior changes. More standalone context. Fewer follow-up clarifications. Your manager stops asking for clarification. Your team stops waiting for your written response.
And the voice thing, the thing you can't measure but can feel, it's real. The next time your manager or a colleague leaves feedback on your async communication, it'll probably sound different. Because you're not hedging anymore. You're not performing careful English. You're just being yourself in writing.