I've been teaching social studies for eighteen years. The part of the job I actually trained for, being the guy in the room who knows the content, is the part software now handles better than I do at 7:45 in the morning. That bothered me for a while. It doesn't now. What replaced it turned out to be harder and more interesting, and I'd rather show you what it looks like with a real lesson than talk about it in the abstract. **The stack on my desk Unit 1, Lesson 1. We called it America's Founding Promise. Students read two primary sources, John Winthrop's 1630 sermon and the preamble to the Declaration, and worked through a set of guided notes that ended in an exit slip. Name the promise. Then name who got left out of it and what that cost them. Twelve of my thirteen kids turned in a sheet. For most of my career, that stack sat on my desk overnight. I'd flip through it, get a general feeling for how it went, and teach the next lesson off that feeling. My feelings about my classes are usually pretty accurate. They were never specific enough to act on. **A report instead of a feeling Now the stack gets scanned and read against the answer key I wrote, then graded on a proficiency scale instead of a percentage: Mastery, Proficient, Developing, Beginning. What comes back is a class content report. Every student has a level, every level has a one-line note about what that particular sheet showed, and there's a summary of where the group stands as a whole. That day it came out 3 Mastery, 8 Proficient, 1 Developing, nobody at Beginning. Good floor. Nobody lost. The distribution isn't the part I use, though. This is. The class was strongest on the hard part. Almost everybody named at least two groups the promise excluded, and a good number gave real consequences instead of vague ones. Forced unpaid labor. No vote. Land taken. That's the move the whole lesson is built around, and they made it. Two things didn't land. About half the class stopped at "all men are created equal" and never got to the rest of it, the claim that those rights don't come from the government in the first place, so the government has no business taking them, and that the whole arrangement runs on consent. The second miss was Winthrop. Students noticed the phrase "city on a hill" and could tell me where else it shows up. Almost none of them caught what he was doing with it, which was taking a religious ideal and hanging it on a brand new country.

You can now book me to speak at your event regarding using AI in education! Let's be honest, you could just contact me through Skool and that will work too. However, now I have a booking website that let's you or someone you know that may be interested to contact me regarding the event. And, of course, I built all of this in about 20 minutes using Claude and hosted it for free with Netlify. https://jeffpeterson.netlify.app/

I've never been on a podcast before. Probably because podcasts are for experts and entertainers. However, I fooled the good people at Beyond the Backpack into thinking I was one of those things. Check it out! https://www.youtube.com/watch?v=jasPeoPC7LU

I used to dread the post-unit data review. Not because the data wasn't useful — it absolutely is. But because pulling it together meant digging through individual reports, cross-referencing spreadsheets, and trying to hold a mental picture of 30+ students across multiple objectives. By the time I had the full picture, half my planning period was gone. That changed when I used Claude Cowork to turn my Chapter 17 Evidence of Learning reports into a live dashboard. -- What Is Cowork, Exactly? Cowork is Anthropic's desktop AI tool — think of it as Claude with hands. It can read your files, run tasks, and build things on your computer, not just in a chat window. I pointed it at my EOL data and asked it to build me something I could actually use. What came back stopped me cold. -- What the Dashboard Shows The artifact Cowork built surfaces everything I care about in one place: - Class-wide mastery rates by objective - Individual student performance across the unit - Gap patterns — which skills need reteaching vs. which are solid - Trend indicators so I can see where students improved or struggled most No more hunting through individual reports. No more mental math. "No more searching for specific reports — now my custom data is live at my fingertips." -- The Part That Changes Everything: It's Live Here's what makes this different from a one-time export or a static chart: the dashboard updates. Whenever I have new EOL data — a new unit, a re-assessment, a mid-unit check — I tell Cowork to refresh, and the dashboard reflects it. It's not connected to a cloud database. It's connected to me and my workflow. I control when it updates, and it updates completely. That means this isn't a tool I use once. It's infrastructure I build on. -- What This Means for Lesson Design and Assessment I'm still unpacking the implications, but here's what I'm already thinking about: Faster feedback loops. If I can see gap data the same day I run an assessment, I can adjust my next lesson before the window closes. That's a fundamentally different planning rhythm.