· Hands-On Labs · 2 min read
Lab #3 - Clean Up Messy Data with 'Edit in Copilot'
The third hands-on lab in my community resource series gives you a deliberately wrecked spreadsheet and walks you through fixing it with Edit with Copilot - one problem at a time, then all at once. Includes a practice file you can download and break yourself.
Edit with Copilot in Excel has been a wildfire at Huron. The feature, recently renamed from Agent Mode, lets Copilot work directly in your spreadsheet: cleaning data, building formulas, creating charts, reasoning through multi-step problems while you watch. Since it rolled out, the interest hasn’t been “oh that’s cool.” It’s been “how do I learn this yesterday.” People are sending me screenshots of their own messy data asking if Copilot can fix it. The answer is almost always yes, but the skill isn’t in the tool. It’s in knowing what to ask for.
That’s what Lab #3 is about. “Clean Messy Data with Copilot in Excel” gives you a purpose-built disaster of a spreadsheet — merged title rows, three date formats fighting each other, 19 category variations for 5 categories, amounts stored as text strings with dollar signs, hidden duplicates, and a hand-typed ”~$28,000??” total in red italic at the bottom that turns out to be off by over $15,000. The file is realistic because it’s modeled on the kind of thing that actually lands in your inbox when someone says “can you take a look at this?”
The lab walks you through fixing each problem one at a time with Edit with Copilot: structural cleanup, date standardization, category normalization, amount conversion, duplicate detection, and summary generation. Then it hands you a fresh copy of the same file and asks you to do it all in one prompt. That second pass is the point. The step-by-step work isn’t busywork — it’s how you learn what to ask for when you’re ready to write the comprehensive prompt. Specificity comes from experience.
This is a different kind of lab from the first two. The Bio Builder was about creating an agent. Copilot Prioritize was about configuring a feature. This one is about developing a skill — the ability to look at messy data, diagnose the problems, and write prompts that fix them. It’s the lab I wish I’d had when people first started asking me for help with their spreadsheets.
If you’re running adoption programs, this one comes white-label ready as always, Excel file included. Go clean something.
Enjoying the read?
I send a short email once a month — behind-the-scenes notes, honest takes, and first word on workshops. No spam, no fluff.
