How I Use AI to Turn Scattered Notes Into a Clear Plan
I rarely have a shortage of ideas. I have the opposite problem. My notes are scattered across too many places to ever add up to a plan on their own.
There are notes in Notion, half-finished thoughts in Google Docs, things I typed into my phone because I was afraid I’d forget them, and occasionally a sentence sitting in a browser tab that made perfect sense three days ago and means absolutely nothing to me now.
The problem isn’t capturing information. I’m very good at capturing information. The problem is turning all of it into something I can actually act on.
More scattered notes did not make me more organized
For a long time, I treated saving information as if it were the same thing as organizing it.
It isn’t. A folder with forty notes is technically organized, in the sense that the notes are all in one place. But if I open that folder and still have to read everything again before I know what matters, I’ve mostly just moved the mess somewhere prettier.
This became especially obvious when I was working on more than one project at the same time. I could have all the information I needed and still spend the first part of a work session trying to reconstruct what I had been thinking when I saved it.
That reconstruction time is what I wanted to get rid of.
What I use AI for now
I don’t ask AI to decide what I should do. I use it when I already have the raw material but can’t see the shape of it yet. I’ll take a messy group of notes and ask it to separate three things: what I already know, what still needs a decision, and what the next actions appear to be.
That distinction matters because a pile of notes usually contains all three mixed together.
One sentence might be a fact. The next is an idea. The next is something I should do tomorrow. Then there’s a question I haven’t answered yet, followed by something I copied from somewhere and no longer remember why I saved.
My brain treats all of those as equal when they’re sitting in the same document. AI is useful because it doesn’t have to.
The prompt is less important than the input
I used to think I needed a very clever prompt. The thing that makes the biggest difference is giving enough context before asking for anything. I usually explain what the notes are about, what I’m trying to achieve, and what I do not want the AI to do. Then I paste the mess underneath.
Something along the lines of:
I’m trying to figure out the next steps for this project. Do not add new ideas yet. First, organize these notes into what is already decided, what is still unclear, and what requires an action from me. That instruction, do not add new ideas yet, turned out to be surprisingly important.
AI is very willing to be helpful. Sometimes too willing. If I give it a messy page and simply ask what I should do, I often get five additional suggestions I never asked for, which is exactly the opposite of what I needed. I already had too much information. The first job was subtraction.
What happens after the AI gives me an answer
I don’t keep the output exactly as it comes back. I read it once, delete anything that feels invented or overconfident, and reduce the action section until there are only a few things I could realistically do next. Then I put those actions back into whatever project page they belong to.
This is the part that makes the process useful for me. AI does the first pass of sorting. I still make the decisions. It is less exciting than asking an AI to run my entire life, but it is also much more reliable.
Where this still breaks
The system works badly when the original notes are bad. If I’ve saved a sentence with no context, AI doesn’t magically know why I cared about it. Sometimes it confidently creates a meaning that was never there, which is a useful reminder that organizing incomplete information does not turn it into complete information.
It also doesn’t solve the problem of capturing too much in the first place. I’m still capable of saving fifteen things when three would have been enough. But at least now I have a way to get from “I know I wrote this down somewhere” to “here is what I actually need to do next” without rereading my entire digital history first.
And for me, that is where AI has been most useful so far: not replacing the thinking, but clearing enough space for me to do it.
