Once you see it, you can’t unsee it
Behind every tedious task is an AI shortcut. Hear Justine's story of vibe-coding her way to a permanent time win.
There’s everyday work, and then there’s work that simply has more moving parts.
Let me tell you about Justine Petrey-Juarez. Earlier this year she was an EA in a family office. The part of her job that stretched her the most was a detailed review process that involved lots of information, lots of names, and plenty of cross-checking. It was a really manual process… and also one that’s critical to get right.
It required cross-referencing email content, Excel spreadsheets and CRM records to make sure there were no overlaps or duplicated relationships. It was intricate, easy to get lost in, and absolutely essential to get right.
I want to share Justine’s story because I personally found it so inspiring to see how she used AI to become a builder. Even though she'd never coded before, she was able to vibe code a Python script to automate this tedious task.
And here's the thing: in the age of AI, we can all be builders. That's incredibly powerful and freeing.
Side note: vibe coding is when you describe what you want your app to do in plain English, and get your AI chatbot to write the actual code for you – you focus on the big picture while AI handles the technical details.
📆 Join this free Carve Shift to hear Justine's story firsthand
Thurs, Jan 22 · 5pm UK · 12pm EST · 9am PST
You'll hear an honest, behind-the-scenes look at what it really takes to make vibe coding work as an EA – the wins, the workarounds, and the lessons learned along the way.
How it started
Justine was in the spring cohort of Carve AI, and midway through the course she started experimenting with ChatGPT to see if it could handle this detailed review process.
She tried feeding it emails. She tried giving it the spreadsheet. She tried explaining the CRM logic.
Short story: it couldn't handle the full workflow.
The model would act like it was working through the challenge, but grind to a halt. So after a few frustrating tries, Justine asked ChatGPT, "What’s the most reliable, secure way to automate this?"
The answer: Python.
Justine had never written a line of Python in her life. Still, she downloaded it, installed libraries like Pandas, and with ChatGPT’s help began vibe coding.
The result
It took her four hours to build something functional. Four hours of painful trial and error, debugging, and feeding ChatGPT tiny chunks of the problem until it spat out workable code.
But once it was done, the workflow was seamless:
scan email files
extract the relevant names or identifiers
compare them against her excel data
cross-check those results against CRM entries
and generate one clean spreadsheet showing any overlaps, complete with references to the exact email where they appeared
She went from hours of manual scanning to a 30-second process: drop the email files in a folder, click Run Me, and get a perfectly formatted spreadsheet showing any conflicts, complete with references to the exact email where they appeared.
ROI? Four hours invested to save 28 hours a year.. forever.
“Now I look at everything through this lens: what’s manual, repeatable, and fixable? Once you see it, you can’t unsee it.”
Justine’s now brought that mindset to a new role working in investor relations. One of her first self-set quarterly goals? Identify and fix three manual processes in her department 💥
4 learnings from Justine's story
1. Start with 'There must be a better way'
The critical shift Justine made wasn't to vibe code using Python – it was the mindset to ask "how could AI help me here?".
Especially when you reach a point where you know your energy could be put to better use.
2. Then pause and let yourself go down the rabbit hole
Even just for 60 seconds!
The real breakthrough isn't in the asking – it's in the pausing ⏸️
Most of us keep moving. We're busy, we have deadlines, we stick to what we know works.
But train yourself to actually pause. Open your AI chatbot and ask how it could help here.
Maybe you don't have time to build the solution right there and then.. so park it. Start a list of tasks you could automate, and chip away at them when you have the headspace.
3. Be intentional in choosing your automation target
Not every annoyance is worth automating. Justine didn’t sink hours into a task that only took her 10 minutes a year. She focused on something high-frequency and high-effort.
Before you start building, consider the annual time cost:
How long does the task take each time?
How often do you do it?
If you can save dozens of hours a year with a one-off investment of a few hours, it's a no brainer. Automate it.
4. Start small, then scale
Justine didn’t start building complex 8-step automations on day one... she started with a contained, easy-to-test process.
Why this works:
You get a quick win (and confidence boost)
You build proof of value you can show your exec
You avoid burning weeks on a complex build that might stall
Once you’ve shipped one, repeat the cycle: notice → pause → assess → automate.
Over time, you’ll build a portfolio of wins that makes you the go-to problem-solver in your team.
Are you thinking about 2026 goals yet?
Close out the year with intention (and a little celebration). Join me and EA coach Jodie Mears for a reflective, practical session on career goal-setting including AI tactics to help you plan 2026 with confidence.
This Thurs, Dec 18 · 6pm UK · 1pm EST · 10am PST