5 Questions to Ask Before You Adopt Any New AI Tool in Your Class
Today I want to hand you the thing I wish someone had handed me before I opened my 7th browser tab this month, each one a different “revolutionary” AI tool a colleague swore would change my life.

Anita Chen

Holi, Soy Anita.
Last time, I told you that most people camp out on layer 2 of Educational Technology - The Tools, and skip the rest of 4 layers around it. Today I want to hand you the thing I wish someone had handed me before I opened my 7th browser tab this month, each one a different “revolutionary” AI tool a colleague swore would change my life.
Here’s the quiet truth nobody puts on the landing page : the number of AI tools you’ve tried is not a metric of progress. It might just be a very expensive form of procrastination dressed up as productivity.
So before you add one more tool to the pile, run it through these five questions first.
Quick answer, if you’re skimming: before adopting any AI tool for teaching, check whether it actually saves you time end-to-end, whether it protects your and your students’ data, how steep the learning curve really is, whether it fits how you already teach, and whether you’d still use it in a month once the novelty wears off. If a tool fails 2 or more of these, it’s not your tool yet, no matter how good the demo looked.
Does it save time, or does it just move the time somewhere else?
A tool that generates a worksheet in 10 secs but takes you 40 mins to fact-check, reformat, and translate hasn’t saved you anything. It relocated the work, not removed it. Before you adopt anything, ask what the full time cost looks like generation, review and cleanup included, not just the flashy part in the demo video.
What happens to your data, and your students’ once you hit submit ?
This one is easy to skip when a tool feels friendly and free. But “free” and “no cost” are the same sentence. Before you paste a single student’s name, a grade, or a piece of their writing into any AI tool, know where that information goes, how long it’s kept, and whether the platform was actually built with education’s privacy standards in mind - not just retrofitted with a privacy policy nobody reads.
How steep is the learning curve, honestly?
Not “how steep does the onboarding video make it look” Instead, it’s “how steep is it at 9pm on a Tuesday, tired, with no one to ask. A tool you need a certification to operate is a tool you’ll open twice and then quietly abandon. The best tools in your stack should feel less like software and more like a colleague who already knows what you’re trying to do.
Does it fit how you already teach ? Or does it ask you to become a different teacher?
This is the one that gets skipped most, and it’s the one from layer 1 of the whole EdTech framework : design comes before tools, not after. A tool that’s genuinely excellent for a flipped classroom is genuinely useless if you don’t run a flipped classroom. Before adopting, ask: does this slot into my actual teaching or does it quietly require me to redesign my teaching around it ? One of those is an upgrade. The other is a part-time job you didn’t apply for.
Would you still be using it in a month - after the novelty wears off?
New tools have a honeymoon period. Everything feels efficient for the first week, partly because you’re paying close attention and partly because it’s new. The real test isn’t week 1. It’s weather you’ve reached for it, unprompted, 3 weeks from now when the excitement has faded and it’s just another Tuesday.
Not every tool needs to become permanent. Some are meant for one unit, one project, one season. The problem is only when you treat every trial like a marriage.
The actual point of this checklist
None of this is about becoming suspicious of AI, or turning into a teacher who “doesn’t trust technology.” Actually, it's the opposite. It’s about spending your limited attention on the two or three tools that actually earn a permanent seat in your workflow, instead of your energy leaking out across twelve tools you tried once and forgot about.
This is the filter I run everything through before I build it into my own AI Agent Toolbox. If a tool can’t survive these five questions, it doesn’t get a permanent home in my process, and it shouldn’t get one in yours either.
I’m going deeper on this, including the actual tools that passed my own five-question test. Over in the newsletter, alongside The Shine Lab. If this is checklist saved you from opening tab number 8, that’s where the rest of it lives.
[Subscribe to the newsletter →]
— Anita
About
Featured Posts
Explore Topics













