Three channels, one place
I've written about how everything I say ends up as text. I've written about why your meeting summary is worthless. But I haven't written about the practical question that determines whether any of it actually works: how do you capture everything?
The answer isn't a single tool. It's three channels that all land in the same place.
Online meetings
Most meetings today happen in Teams, Zoom, Meet. Teams has transcription built in — you can copy the text directly after the meeting. It works. There's a bit of manual effort, a bit of copy-paste, but it's entirely viable.
For those who want something more streamlined, there's Fireflies. It joins your meetings, records them, transcribes them. And the important part: it has an API. I pull all my Fireflies transcripts centrally every fifteen minutes. Automatically. No manual step. The meeting ends, and within a quarter of an hour the transcript is ready to be processed.
It's not magic. It's an integration and a scheduled call. But it eliminates the friction that makes people stop capturing their meetings after the first week.
In-person meetings
Not every conversation happens in front of a screen. Client meetings, workshops, lunches where the important things are said between courses — those need to be captured too.
I use record-me on my iPhone or iPad. Every in-person meeting gets recorded. The audio file is then passed through the same pipeline as my Fireflies meetings. Same place, same processing, same chain afterward. It doesn't matter whether the meeting happened in Teams or in a conference room — the transcript ends up in the same spot.
Newer versions of iOS include Apple Intelligence, which lets you copy text directly from a recording. That simplifies things further. But the principle is the same: in-person conversations can't be treated differently than digital ones. If you only capture online meetings, you're losing half of what's actually said.
Phone calls
The third channel most people miss entirely. Inbound and outbound calls — the ones that often contain the fastest decisions, the most unfiltered information, the things people say on the phone that they would never have written in an email.
I use Sonetel as my phone service. They have a call recording feature and an API for retrieving the audio files. Same principle as Fireflies — automatic retrieval, centrally stored, ready for transcription.
It's the channel people forget most often. And it's the channel where the most valuable information tends to live, precisely because phone calls are spontaneous, unprepared, honest.
The convergence point
Three channels. One place.
Online meetings via Fireflies. In-person meetings via record-me. Phone calls via Sonetel. Everything lands in the same pipeline. Everything becomes transcripts. Everything becomes searchable, structured text.
It's not three separate systems. It's three inputs into the same system. And that's the point — it doesn't matter where the conversation happens. It gets captured. It gets processed. It doesn't disappear.
Then the chain takes over. Preparation for the next meeting based on what was said in the last one. Summaries, decisions, follow-ups. Project tracking if needed. Daily overview. All the things I've described in the other articles — but it starts here. With the audio actually being captured, regardless of channel.
The only thing required
The infrastructure isn't complicated. It isn't expensive. Fireflies has reasonable plans. Sonetel too. record-me is a tool I built myself, but there are alternatives.
What it takes is the decision to actually do it. To commit to the idea that no conversation is too unimportant to capture. To stop treating in-person meetings and phone calls as exceptions to the system.
Most people who say they work with AI-assisted workflows capture their Teams meetings and nothing else. That's one third of the picture. The rest disappears.
Nothing needs to disappear. The infrastructure exists. The channels exist. The only thing missing is the habit.