AI voice dictation for work: Wispr Flow + Claude.
I use Wispr Flow to talk instead of type, and Claude to turn what I said into something finished. Flow lets me dictate fast into any app - prompts, emails, notes, briefs - so the raw thought lands in seconds. Claude then structures it, sharpens it, and puts it in my voice. I use the pair for basically everything: dictate a messy brain-dump, hand it to Claude, get back a clean email, sequence, ICAP or post. Speaking is faster than typing and far faster than writing well - so I speak the thinking and let Claude do the writing.
How I use this, and what for
I use Wispr Flow to talk instead of type, and Claude to turn what I said into something finished. Flow lets me dictate fast into any app - prompts, emails, notes, briefs - so the raw thought lands in seconds. Claude then structures it, sharpens it, and puts it in my voice. I use the pair for basically everything: dictate a messy brain-dump, hand it to Claude, get back a clean email, sequence, ICAP or post. Speaking is faster than typing and far faster than writing well - so I speak the thinking and let Claude do the writing.
What the AI actually does
Two kinds of leverage, stacked.
Wispr Flow takes your speech and turns it into clean text directly where your cursor sits - Gmail, Slack, a form, the Claude prompt box, anywhere. It runs the speech through AI that strips filler words, fixes sentence structure, and formats for the app you are in, so you capture a full thought at the speed you think it. The makers put dictation at roughly 3-4x typing speed, with users reporting clean output in the 150-220 words-a-minute range.
Claude takes that fast-but-rough transcript - rambling, unstructured, not on-brand - and turns it into the finished thing: structured, tight, in your voice, brand-checked. You think out loud, and what comes out the other side is professional.
The tools and APIs you need
The right prompts
Dictated brain-dump to finished asset (speak the whole block)
```
Right, I just got off a call with Acme, they liked the product but they are
worried about ramp time and whether juniors can hit quota fast enough, the
champion is the VP Sales but the blocker is finance... take all of that,
clean it up, and turn it into a follow-up email in my voice plus three
bullet next-steps I can drop into the CRM.
```
Rough idea to polished post
```
Quick thought - companies keep hiring SDRs on confidence and then they
churn in 90 days because nobody checked for coachability... turn that
into a LinkedIn post in my voice, 150 words, one sharp point and a hook.
```
Set Flow up to stop fighting you
```
(In Wispr Flow's dictionary) add: meritt, ICAP, SDR, AE, CSM, Lauren,
MEDDIC, Apollo, Common Room - so it stops mis-transcribing the words I
say fifty times a day.
```
How to do it
Dictate the raw thought into Claude
Open the prompt box, hit Flow, and just talk - the full messy version. Do not self-edit while speaking; that is Claude's job.
Tell Claude what to make of it
End the dictation with the instruction: "turn that into a cold email", "structure that into an ICAP", "tighten that into a LinkedIn post in my voice".
Let the right skill do the shaping
Name the skill so voice and brand are baked in - `will-voice` for emails and posts, the blog skills for content, `meritt-brand` for a brand check.
Review, do not rewrite
Read it, nudge it with another quick spoken instruction ("more direct", "cut the second paragraph"). Iterate by voice, not by keyboard.
Use it as the front door to every other lesson
Prospecting, call review, CRM logging, newsletters - all of them start with you saying what you want. Flow gets it in; Claude gets it done.
See the difference
Doing it by hand: ai voice dictation for work: wispr flow + claude the manual way - slow, and the first thing to slip when you are busy.
With AI: you describe what you want in plain English and it does the work, on-brand, in minutes.
Let AI carry the heavy lifting; you keep the judgement and the final say.
How you'll know it's working
- You capture a full thought at speaking speed - no keyboard between you and the idea.
- What comes back is finished and on-brand, not a tidied-up transcript.
- You edit by talking, not by typing.
- Every other skill in this playbook starts here: speak it, and Claude ships it professional.
Questions people ask
Is voice dictation actually faster than typing for work?
For most people, yes. Speaking runs roughly 3-4x faster than typing, and [Wispr Flow](https://wisprflow.ai/) cleans the speech as it goes, so you are not just talking fast - you are capturing a usable draft at the speed you think. The win is biggest on first drafts, where the bottleneck is getting the thought out, not perfecting it.
Will it mis-hear sales jargon and names?
At first it might, which is why the dictionary matters. Add your recurring terms - meritt, ICAP, SDR, names like Lauren - and Flow stops mis-transcribing them. With [auto-add to dictionary](https://docs.wisprflow.ai/articles/4052411709-teach-flow-your-words-with-the-dictionary) on, it learns a word the first time you correct it.
Do I still need Claude if Flow already cleans the text?
Yes - they do different jobs. Flow gives you a clean transcript of what you said; it is still a spoken brain-dump, not a finished email. Claude is what structures it, tightens it, and puts it in your voice and brand. Flow gets the input in fast; Claude gets the output out professional.
Hire with Assessment.
£7-10k flat fee. The methodology, delivered.
See Hire with Assessment