Back in June I wrote about building a personal operations system with Claude Code. I got asked this week how it’s going, so here’s an update. TL;DR: it’s a work in progress and I’m still making tweaks.
When I started it, I thought I wanted a tool to show me everything I had going on: every Slack mention, every meeting, and every email. It worked… way too well. The list grew faster than I could clear it only to find it stale again a week later. So essentially, I felt like I was working for the tool and it was an extra task on my list, which is exactly NOT what I wanted.

In my current role, there are things I just need to be aware of but don’t need an action on my part, so I tried sorting things into buckets: what I need to act on and what I just need to know. I’d have Claude run that every morning so I could see what was on my list and what I should know about. That’s been helping tame the chaos.
Then I noticed Claude was making weird decisions about what ended up on my list. My favorite: it drafted my weekly update saying I was spending three days in New York for a team AI enablement session. I was going to New York, but the event Claude referenced had been cancelled, the dates had changed, and I was going for an entirely different reason. It had found an old calendar entry and confidently built a whole story around it. I caught it, and it wasn’t a huge deal since it was only for my own awareness, but it really highlighted that I’m, at the end of the day, the editor.
Where I am now: every morning my morning brief routine reads my calendar, meeting notes, Slack, and sent mail fresh each time. I archived the old giant list and started over with a shorter one I picked myself. I get a short message that’s gotten much better at highlighting the top things I need to ship that day, and who’s waiting on me. The goal is to wrangle all my different focus areas into one place, and it’s getting better at that. It’s also gotten better at seeing patterns and making connections between them.
What I didn’t expect
Along the way I got better at using Claude for everything else. I think through hard problems with it, especially when there are a lot of sources that I need to pull together, which may take some deep research and analysis. I build things I used to wait on someone else for: a CRM that flags updates from my email and Slack every morning, a community events dashboard to see what’s happening across a huge global landscape, and even migrated a website for a non-profit I work with from another platform over to WordPress. I have it check work against rules I’ve set, like reviewing invoices against contracts before I approve them. My memory files are shared across Claude Code, Cowork, and chat, so whichever one I open already knows my people and my projects.
Overall, it’s working. It started as a way to keep track of my work, and it’s turned into the place where a lot of my work happens. Let’s see where I’m at in another 5 months!




































































































