Found a workflow I keep coming back to. When you're in a complex conversation, whether it's a strategy session or just working through a problem with people, the ideas are covered, but agreement may fall short and retention of any alignment is a huge challenge the more complex ideas.
What I've been doing is recording it with Voice Memos on iOS and letting it auto-transcribe to text. Then I hand the transcript to Claude and ask it to distill the whole thing into key findings, decisions made, and open questions, key quotes, and next steps. I tell it to lose no nuance due to the need to refer back, but to organize the conversation well.
That part alone has been a game changer. But, then there's a next step that I wasn't expecting to work as well as it does. You can take that distilled output and generate a visual representation of the conversation. A single image that maps how the ideas connect. I use a custom Claude skill that expertly writes optimized prompts for Google's Gemini image model (Nano Banana Pro v3) which is great at text and layouts. I paste the prompt into Google's AI Studio, where you access the Nano Banana Pro model with your API key, and what comes back is way more accurate and useful than you'd think.
The group can look at one image and immediately see the shape of what they discussed. It becomes something to build on next session instead of starting from scratch.
About ten minutes of work after the conversation ends. You walk away with a clean document and a visual, both shareable.
I put together a short reference guide for the full workflow if anyone wants to try it.
Images in this post are an exact example of what this can be. I even used the Voice Memo as a starting point. (The last one was the redirect prompt: "Now do prompt that focuses more on the team communications flow benefits of this workflow mixed into a planning process. ")
