AI is not my favorite. You’ve probably picked up on that if you’ve read newsletters from the last few months. I don’t want AI doing my creative work. I don’t need to use AI for “inspiration” or to write poetry or do write my newsletter.

And yet, I’ve found AI to be very helpful in my work life.

I’ve found it helpful to have a “personal AI policy” and also determine where, when, and how I am using it in my work life.

My Personal AI Policy

• I don’t use AI in any of my creative work, ever. Period. I don’t use AI to write my newsletters, assist in my art making, or create assets for my website.

• I do occasionally use AI to help with tasks like meal prep and planning, when I need help coming up with things I can eat. Since my menu is fairly limited due to histamine intolerance as a result of MCAS, I can’t eat a lot of things unless I want to be miserable. AI has been helpful in this, especially Claude as it will create menus for me and then I can quickly and easily shop for groceries.

• I occasionally use AI to summarize research papers, when I need to know the blood serum level of various compounds to achieve effectiveness, as an example. This helps me translate how much Quercetin I have to take on a daily basis (hint: 3 grams) to have an impact on my MCAS.

• As someone with ADHD, sometimes having a way to map out tasks step by step with AI is helpful for me, especially if I’m having a chronic illness flare.

• Since I use AI, I try to offset my carbon footprint each month. I don’t always set aside the money to do this, but I do try to make room to do this through trusted means beyond just buying carbon credits.

How I Use AI at Work

This month we got access to Claude at work, including Claude Co-work and Claude Code. I’ve found a lot of value in both, and I want to share how I use these tools at work.

• Again, I’m not interested in having Claude help me do the more creative aspects of my job, like customer research, literal design, or figuring out what should come next in the sequence of things we’re building.

• However, I do find it incredibly helpful to take over a lot of the mundane things that take up a lot of time. Here are some examples:

  • I’ve used Claude Co-work with the Microsoft365 Connector to automate as much of my calendar as I possibly can. It can find and fix situations where I’m double- or triple-booked. It gives me a heads up on meetings that are coming up. It will also help me find opportunities to maximize or optimize my time.

  • I’ve used Claude Co-work and the Jira connector to add context to acceptance criteria, build out epic’s worth of tickets, help sequence work, and generally get prepped and ready for decomp - mostly without me actually needing to open Jira other than to update the order of the tickets in the backlog.

  • I’ve used Co-work to clean up my inbox on a weekly basis, including helping me archive emails in the appropriate folders using the Google Chrome MCP, which basically then allows Outlook to run in Google and allows Claude to function by clicking on various things in the browser. You can follow along and watch what it’s doing, which I did the first few times to ensure it wouldn’t delete important things, like training emails or deadline related things.

  • Data analysis - Claude is great at doing data analysis with supervision. We’ve set up the Mixpanel MCP. I’m interested to see how it does at data analysis through an MCP using a data connector.

  • The one shortcoming I’ve found with cowork is that it really doesn’t share context across chats, which can make it a little tricky to get the results without a lot of context sharing. I’m working on building a markdown file I can add to each and every cowork chat so that it starts from that same shared context, but that feels a bit tedious to be honest.

I’d love to hear how you’re using AI - or not using it! Let me know. Love it? Hate it? Somewhere in between like me?