Phase 1: Capture
Raw material: Transcription is the foundation where everything begins.
The principles give direction. Here it becomes concrete: a collection of what becomes possible once you start recording conversations.
Where are you?
You've had a conversation. A workshop, an interview, a session with people discussing something important.
Until now, that was fine. You had your notes, your shared memories. But the exact words, those you didn't have.
Something has changed. We can now make transcripts. Fast, cheap, good enough. And that raises a question: what can you actually do with that, in a social setting?
That's what this field guide is about.
The story: the bicycle helmet
Friend, facilitator and father of two kids, Maarten organized a parent-teacher evening about smartphone-free growing up. As the host, he guided the conversation, asked questions, held the energy. He knew what he wanted to achieve, and he was curious what AI would do with the transcript.
After the evening, he had the recording transcribed and analyzed by AI. One quote stood out:
"Your kid goes to school by bike for the first time, helmet on. Comes home and says: 'Nobody in my class wears a helmet. I'm not wearing one either, or I won't fit in.'"
Maarten: "This is exactly the core. This captures what we as parents are wrestling with."
The quote reinforced what he already knew. Now he had concrete words to open the next session with, to let others recognize what it's really about.
This is one thing that transcription with AI makes possible. There's more.
What can AI do here?
Transcription: the raw material
Transcription is raw material. Foundation for all kinds of things: capturing collective wisdom, uncovering patterns, checking your own intuition, growing as a facilitator.
The barrier has dropped to almost zero. An hour of audio costs a few minutes and little money, or even free if you do it locally. And it's not just about conversations with others. With dictation, you also capture your own thoughts: you speak, stop, and within seconds everything is there as text.
But the question isn't how you get text. The question is what you do with it. AI can reread the entire conversation and find things you missed, not because you weren't paying attention, but because you're human. Recognizing structure: what was decided, who does what, which questions remained open?
Now you have text. The question becomes: how do you use what's in there?
Language as ownership
This is one of the reasons why transcription is so valuable.
The difference between "you're talking to a wall" and "communication issues" seems small. But the first is what someone said: raw, emotional, recognizable. The second is consultant-speak. And that difference determines whether people recognize themselves in the output.
When people see their own words reflected back, ownership emerges. Not because the synthesis is perfect, but because it's recognizable. "Yes, that's what we said": that's the moment where participation turns into commitment.
This is the third principle from the six principles: preserving their words verbatim isn't a detail, it's the mechanism.
Adopting the style of a source document
Sometimes the transcript doesn't just need to capture participants' language; it also needs to fit into a format that other (external) parties will accept. A transformation plan for the health insurer, a vision document for the municipality.
There's an approved document. The question: how do you write new content in the same style, while preserving participants' words?
AI can analyze the style characteristics of a source document and apply them to new information. Structure, tone, terminology; systematically adopted. The form may change, as long as the ownership stays intact.
What stays human?
This is the question that runs through this entire field guide. As AI becomes more intelligent, what we do shifts, but doesn't disappear. In the six principles this kept coming back: what can AI do, and what stays ours?
With transcription, the answer is clear.
Transcription is raw material, not an end product. AI can record what was said, but we determine what matters. Which quotes do we share? Which patterns do we emphasize? How do we move the conversation forward?
And perhaps more importantly: the safety in the room, sensing when someone needs space, reading faces. That stays human work. AI can analyze, but feeling the moment, it can't do that.
| AI can... | Human must... |
|---|---|
| Record everything | Determine what matters |
| Quote verbatim | Sense whether people recognize themselves |
| Add structure | Check whether it matches the atmosphere |
| Recognize style characteristics | Judge whether it fits the context |
Tensions in this phase
These are the choices that keep coming back when working with transcriptions. Not as pitfalls to avoid, but as trade-offs you keep making anew. On the technique pages you'll find more specific tensions per technique.
Verbatim vs. cleaned up The urge to make everything "neat" and professional is strong. But the power often sits in the raw words. "You're talking to a wall" hits differently than "communication issues." The tension: when do you preserve the rawness, when do you make it accessible?
Recording vs. using The barrier for transcription is so low that it feels natural to record everything. But recording is not the same as using. Without intention, transcripts gather dust. The tension: when do you transcribe with a purpose, when do you archive for later?
Convenience vs. privacy The tools are so easy. Upload, click, done. But not every conversation belongs in the cloud. The tension: when do you choose convenience, when do you choose local? And what does "safe" actually mean for you and the people you work with?
Remembering everything vs. trusting the transcript You don't need to be a superhero. You're allowed to miss things: that's what the transcript is for. The art is being in the moment, knowing you can go back and read later.
Getting started
Want to get started right away? The technique Transcription as foundation lays the groundwork, from recording to your first analysis.
Techniques in this phase
| Technique | What it does | When to use |
|---|---|---|
| Transcription as foundation | Record, recognize structure, make searchable | For any conversation rich enough to preserve |
| Language as ownership | Create ownership through verbatim quoting | For syntheses and feedback to participants |
| Source document style cloning | New documents in existing style | When you need to match organizational standards |