In today’s Everyday AI video, I take you on a journey with me to set up and execute two tasks in Claude Cowork.
One was easy. The other was a bit of an experiment.
Both saved me real time — and gave me a few things worth sharing.
TASK 1: The participant reconciliation that used to take me an hour
Every time I run an AI Deep Dive Lab, there’s an admin headache that comes with it.
Participants change sessions. Others withdraw at the last minute. A flood of back-and-forth emails lands in my inbox, and someone has to reconcile all of that against the original participant list.
That someone used to be me. Or someone on my team. Squinting at an Excel spreadsheet, cross-referencing emails, manually updating names and sessions.
Yesterday, I handed the whole thing to Claude Cowork.
I gave it my participant spreadsheet and told it to look for emails about the course. That’s it. No pointing it to a specific thread. No digging through my inbox first. Just a plain-language instruction, and Claude found the relevant emails itself.
Three minutes later, it handed me a clean, reconciled list — correctly accounting for every change and withdrawal.
Three minutes.
👇
I’m not sharing this to impress you — you’ve seen demos before. I’m sharing it because there’s a difference between watching a slick vendor demonstration and watching someone use AI on an actual Tuesday afternoon admin task.
This is what the usage-to-value ladder looks like in practice. Not a Fortune 500 case study. Just a small, annoying, recurring task — handled.
TASK 2: The annual headache I vibe-coded away
This one’s a bit more interesting.
Every year I facilitate a leadership program that includes strengths profiling. And every year, the same thing happens: I spend more time than I’d like tracking who’s completed their profile, chasing stragglers, and then scrambling to pull completed profiles into one folder so I can hand them out to participants who show up to the debrief without theirs.
It’s not a big job. But it’s fiddly, repetitive, and happens every single year.
This year I decided to fix it properly.
I used what people are calling “vibe coding” — describing what you want in plain language and letting AI build the automation for you. No prior coding experience required. Just a clear description of the problem and a bit of back-and-forth with Claude.
The result? An automated task that monitors who’s submitted their profile, keeps a live tracker updated, and drops completed profiles into a shared folder — ready for the debrief session.
👇
Here’s something I want you to notice in that video: I’m sitting there watching Claude work through the task in real time, because I was filming it.
In real life, I wouldn’t do that.
I’d kick it off and go make a cup of coffee. Take the dog for a walk. Get on with other work. Claude runs in the background and pings me when it’s done.
That’s what agentic AI actually means — not AI that assists you while you watch, but AI that works while you’re not in the room.
WHAT I WANT YOU TO TAKE FROM THIS
Most of the conversations I have with leaders are about AI at scale — strategy, governance, workforce transformation.
That stuff matters. But it can wait five minutes.
Because if you haven’t started using AI to remove friction from your own day, you’re asking your teams to do something you haven’t done yourself.
Both of these tasks were experiments. Neither was high-stakes. And both saved real time that I’ll now spend on something more valuable.
The question I’d encourage you to sit with this week:
What’s the thing I do repeatedly that I resent doing? Could AI do it instead?
What’s the recurring task in your week that you haven’t tried handing to AI yet?
Hit reply and tell me — I read every response.
Until next week, Kim
P.S. I’m running public AI Deep Dive Labs in July and August for specific functions (EAs, Engineers, Mortgage Brokers, and a general, all purpose catch all).
I’m also taking bookings for in-house labs. Reply to this email for details — and if you’re interested in an in-house lab and book before Tuesday, I’ll throw in a 10% EOFY discount. DM me.
