I’m teaming up with Holly Clark for a free webinar about student AI literacy!

We’ll be discussing concepts right out of my new book, AI Literacy in Any Class — practical ways teachers can level up their teaching AND integrate small AI literacy lessons for students.

DATE: Monday, May 4
TIME: 8pm U.S. Eastern time / 5pm Pacific time
COST: FREE

There’s no registration, so just set a reminder in your calendar with this link and show up!

BONUS: Holly has already recorded FOUR WEBINARS in this series — and you can go watch the replays right now.

📚 New AI resources this week

1️⃣ What Does Real AI Leadership Look Like in Schools? (via Digital Promise) — An exploration of why modern leaders must act as both "guardians and explorers," focusing on auditing "shadow AI" and building community-wide AI literacy.

2️⃣ Federal Grants Will Prioritize AI Initiatives in Schools (via K-12 Dive) — A new federal rule signals that AI-related programs will receive increased funding priority, shaping district strategy and grant opportunities.

3️⃣ Is Your School’s Approach to AI Too Flexible? (via Education Week) — A fresh opinion piece argues that relying on “guidelines” instead of policy can create inconsistency and confusion across schools.

Last week’s poll: In your opinion, what's the most powerful way that AI can analyze data in education?

🟩🟩🟩🟩🟩🟩 Classroom data (formative assessment, quizzes) (25)
🟨⬜️⬜️⬜️⬜️⬜️ Community/state/national trend data (6)
🟨🟨🟨⬜️⬜️⬜️ School/district data (16)
⬜️⬜️⬜️⬜️⬜️⬜️ Data gathered from custom surveys (3)
⬜️⬜️⬜️⬜️⬜️⬜️ Other ... (3)

Classroom data: The issue here is that if we don't design these tools carefully first the data we collect and analyze will really be worthless.

School/district data: School and district data to show the importance of data over all classes including electives, CTE, or any classes non-core. How can non-core classes help and why it is important for our students. How to get the non-core to buy into the importance of what and how our students perform. — B. Payne

Other: All of the above (as long as there is no PII 🙂). We can use AI to help identify trends and anomalies, project future data at any level, compare existing lessons and assessments for the best success, and dig down to find some true root causes and solutions in a fraction of time currently spent on the data. — Kerry Fergason

School/district data: Although the analysis of Classroom data is so powerful compared to what was previously possible, I just don't see most teachers investing the time to learn how to do so (even though it can be relatively straight forward). AND, I predict that very few teachers would take the time to "run" the AI analyzer to create the results. As a default, schools and districts might have more capacity to find/run the AI data analyzer and share highlights or concerns with schools and teachers. — C. Kannekens

What’s a topic you’d like to see covered here? Hit REPLY to this email and let me know.

Have you done anything you’d like to share with the AI for Admins community? Hit REPLY and let me know.

Would you like to write a guest post to support and equip AI for Admins readers? Hit REPLY and let me know.

🗳 Poll: AI data upgrades

Instructions:

📈 Using AI to make sense of data in schools (Part 2)

Last week, we talked about what data you can analyze with AI — and the safeguards you need to have in place before you start.

This week, we're getting into the how. Specific tools. Specific prompts. A workflow you can actually use.

Let's dig in.

The single biggest upgrade most people can make when analyzing data is how they prompt AI.

Most people (I’ve done this, too!) dump a spreadsheet into an AI tool and type something like: "What does this tell me?"

That works sometimes. (I think we hope it’s an “easy button.”) But you'll get much better results if you give the AI a role and a focus before you share anything.

Try this framing instead:

Watch the three parts in action below:

Giving the AI context — a role, a data description, a clear goal — is the difference between a generic response and one that's actually useful.

Most admins use AI in descriptive mode. They want AI to answer: "What happened?" You feed it last year's test scores and ask what they show.

The problem? It’s looking back. That helps for certain things, but it can also help you look forward.

AI is also capable of predictive analysis. That answers, "What might happen?" That's where things get really interesting for school leaders.

Feed it two or three years of attendance data and ask: "Based on these patterns, which student groups are showing early warning signs of chronic absenteeism?"

Feed it longitudinal enrollment data and ask: "Based on these trends, what might our enrollment look like in three years — and what should we be planning for?"

Remember: Predictive analysis is NOT MAGIC. It isn’t a crystal ball that tells you the future. But it will show you where the data is trending … and if it follows that trend, it can help you to plan accordingly.

Lots of the tools you might use to do this work won’t be foreign to you — thankfully! It helps to use the right one in the right situation. Here's a quick breakdown:

ChatGPT, Claude, or Gemini — The workhorse for data analysis. Paste in a CSV, a table, or copied data and start asking questions. All three will get the job done … but you might have a favorite based on your preferences or what your school subscribes to. (Reminder from Part 1: check whether you're using a free or school-managed account, and what that means for your data.) ChatGPT (with Advanced Data Analysis) and Gemini (with data tools) have built-in intelligence layers that make data analysis smarter … and there’s nothing extra you need to do to use them.

NotebookLM (Google) — I LOVE NotebookLM when I’m working with multiple documents at once (PDFs, reports, data files, etc.). You can ask questions across all of them simultaneously. The real power is what you can make with those documents. Examples: summaries to share with your team or board, infographics, slides, video summaries, audio summaries, etc. I’ll ask it to make infographics about specific concepts — or to compare multiple concepts. Plus, Google has stated it doesn't use your NotebookLM data to train future models.

Built-in platform tools — Before you export anything, look at what your existing tools already offer. In Khan Academy, Quizizz, Brisk Teaching, and similar platforms, there's often an "Insights" tab that's doing AI-powered analysis automatically. All you have to do is click on it and see what it says. It might have already done the work for you.

Your internal school data tells you what's happening inside your building. But community data — about your neighborhood, city, state, country — can tell you about trends around you that might impact you — or inform your decisions.

The good news: this data is free, public, and more usable than ever. Here are the sources worth bookmarking:

data.census.gov — The gold standard. Income, education levels, employment, language spoken at home — all the way down to ZIP code or census tract level. The American Community Survey (ACS) is updated annually and is especially useful for school planning.

DataUSA.io — Takes the same government data and turns it into clean visual dashboards. Want to know what industries dominate your community — and whether your CTE programs align with them? Start here.

censusreporter.org — Even simpler than DataUSA. Quick comparisons between locations, downloadable data, easy charts. Think of it as census data without the headache.

Bureau of Labor Statistics (bls.gov) — Regional jobs and wages data. Useful for career pathway alignment and understanding what the economic landscape looks like for your graduates.

Download a dataset (or copy/paste a table), give the AI some context about your school, and start asking questions: "What challenges might a school face based on this community data?" or "How might these economic trends affect our enrollment or family engagement over the next few years?"

Remember: AI finds patterns in whatever you give it. That can be good and bad.

If the data you're feeding it already reflects inequities (and a lot of school data does), the AI's analysis might reinforce those inequities rather than challenge them. Disciplinary data is a good example: if certain student groups are overrepresented in referrals, AI might treat that as a pattern to explain rather than a pattern to question.

Be the human in the loop. Always question and critique. Uncover the patters and reasons that AI doesn’t see … things that you notice because you’re human and you’re living there. Ask whether the data itself is telling the whole story.

Copy and paste any of these once your data is ready:

It’s still (thankfully!) up to the humans to make decisions. AI can make you a sharper decision-maker, though. It can help you decide faster, with more context, and with fewer blind spots.

What data have you been meaning to dig into? Hit reply — I'd love to hear what you're working with.