What’s the missing word? Messing? Faffing? Something else? There’s not enough information to know for sure, therefore we fill it in. What influences our fill word? Remember that feeling, because that's what this issue is about.

In education we rarely hear the word "enablement." When we do, it's a warning: that parent is enabling their child. It means doing too much and rescuing too fast, so the child never learns to manage on their own.

In business, enablement means the opposite. It means giving people the tools, knowledge, and confidence to do the work themselves. It's the same word with opposite meanings, and now educators are being asked to "enable AI."

So here's the translation. AI enablement is scaffolding, not rescuing. It means knowing just enough about how these tools work to get good results, on your own terms, and to notice when they're wrong. The ideal feeling would be like you’re walking on pavement, instead of trying to run through deep mud or getting stuck.

Most of what you need fits into three principles.

The quality of what you put in sets the ceiling on what you get out. Give it a vague, half-formed input and it does exactly what you just did with that missing word; it fills the gap with a guess. No prompt, model, or premium plan can rescue a messy input.

It's so simple it could make you cry. And it matters, because most of us are on free accounts or limited by our school or district. Once you know how to put information in well, you need far less computing power than the industry suggests. Remember it is still a business.

When the notes get a client's name wrong or a number off by a zero, the follow-up goes out wrong too. Or someone hand-fixes it first. Mostly right is still wrong.

Wispr Flow Notetaker uses your dictionary and calendar, so names, numbers, product names, and acronyms come out spelled right. It pops up when your meeting starts and captures Zoom, Google Meet, Teams, and Slack huddles with one click, no bot joining the call.

Weeks later, ask one question across every meeting you've had and get the answer with citations. Try Notetaker free on Windows and Mac.

Try Notetaker Free

When educators hear "data," we think of test scores and dashboards. That's too narrow. Data is information, and in a classroom almost everything counts:

Data isn't only output. It's input too, and teaching is one long exchange of information.

Our brains are remarkable at processing all of that at once: reasoning, judging, making connections. AI doesn't think that way. It spots patterns in how humans do things and reproduces them very quickly. It can look intelligent, but for right now it’s simulating intelligence.

This reframe is what makes Principle 1… principle 1. Everything you already handle is the input. Your lesson plans, rubrics, model answers, and student work are what AI works from. That means you already have a lot of control over the quality of what AI gives you.

Never trust AI output as-is. Check it, every time, and especially while you're learning.

Start with easy wins: tasks you'd do anyway, where you already know the right answer. Plan the family menu for the week. Organize your Thanksgiving thank-yous. Put everyone's birthdays into a calendar, then check that it got them right. These tasks are low-risk and easy to check, so you quickly learn where the tool is reliable and where it isn't.

Evaluating output properly needs its own deep dive. More on that below.

Three habits turn Principle 1 into practice.

Clean, uniform data. Give AI tidy, consistent raw material: the same format throughout, no clutter, nothing half-finished. "Raw" here means your source content, like the facts, texts, and questions, rather than a finished worksheet. Raw material is flexible.

Chunk it. It's nearly Thanksgiving, so think about cooking. A whole 12-pound turkey really only gets cooked one way, and slowly. A turkey breast you can grill, roast, slice, or shred. It's the same with a pumpkin. The bigger it is, the harder it is to cook whole. You cut it into chunks, and it becomes soup, muffins, or pie. Your content works the same way. Feed AI a whole book at once and you get a one-off result. Break it into clean pieces, pair them with a template, and the same ingredients become a lesson, practice questions, a parent letter, or a summary.

Text over images. Screenshots and photos of documents take AI longer to read, and they use more tokens. Tokens are the units your plan counts against your limit. Whenever you can, upload or paste text from a Word doc or PDF instead. Plenty of people upload pictures out of habit and spend their allowance on the format instead of the work.

The reward: fewer corrections, longer sessions before you hit your limit, and more of that limit spent on real work!!

Different AI models are good at different things. The basic move is to pick one you enjoy using, learn it well enough to cover most of your needs, and only branch out when you need something more.

Choose based on your own skills and gaps. I talk a lot, I work through a lot of documents, and I have a text disability. I need a model that handles long text well. My first degree is in fine arts, so I don't need a model that makes images. I can sketch what I want myself. A model which handles long-form text well (or long-winded explanations), and doesn’t natively generate images is a great fit as my main tool. Someone less visual and more concise might choose a different one.

Choose tools that multiply what you are good at and support you where you struggle.

That's the agency, and it's the topic of Issue 4.

Live unBoxing, Sunday, November 8. We'll unpack how to evaluate AI output, which is the "how" behind Principle 3. https://luma.com/kw9kjrdw 

The book. My full teacher enablement guide is on pre-sale now for €25 (about $29 USD), and it will be available at the end of November. https://www.thinkoutsidetheblackbox.com/products/totbb-educators-ai-enablement-guide 

Small prompt: What's the messiest thing you've ever fed an AI, and what came back? Or which edtech tools do you use to turn your raw content into materials? Leave it in the comments.

Shit in, shit out. Clean up your inputs and enjoy better outputs. 🎃

Think outside the black box.
Lean. Learn. Lead.
- Jessica (JV)