When to cancel an AI teaching tool
Cancel when the tool stops giving back more time than it costs — not when it disappoints you once. On a $14.95 plan that line is 26 minutes a month, about two resources kept rather than rewritten. Everything below is about finding out where your own usage sits, including the honest answer that for some teachers it is already below the line.
Cancel when
Two months running land under your break-even line — 26 minutes a month on Starter at $35 an hour. Below that, the plan is a donation.
Step down when
You never reach your monthly limit. Moving to the tier below keeps the tool, your saved work and most of the saving, and cuts the part you were not using.
Keep it when
You reach for it unprompted, the edits are short, and you finish planning earlier than you did last term. That is the whole test, and no page should talk you out of it.
Six signals it is time to go
None of these is about quality. A tool can be excellent and still be the wrong thing to keep paying for, and each signal below comes with the check that confirms it in under a minute.
You have not generated anything in three weeks
Logging in is not use. If the last resource you actually generated predates the last unit you taught, the subscription is paying for a good intention. This is the most common and least dramatic reason to cancel.
Check the date on the last thing you generated. If it is more than a month old, you have your answer.
You rewrite more than you keep
The saving on this whole category of tools rests on output being a draft you edit lightly. If you rewrite from the first line, you are doing the planning twice and paying for the first attempt. Cancelling costs you nothing but the attempt.
Look at your last five resources. Count how many you would have been happy to hand to a class with edits under ten minutes.
You only reach for it in the first week of a unit
Usage clusters hard around unit starts and assessment week, then goes quiet for a fortnight. That is a real saving, but it is a few days a month — which is a free-tier shape of usage, not a subscription shape.
Open your generation history and look at the gaps between dates. Two long gaps a month is a downgrade signal.
You are paying for a tier you never fill
The step above the entry plan buys capacity, and most teachers use it in two months of the year — exam season and the start of a course. Paying monthly for headroom you use twice is the quiet version of wasting money.
Check whether you have ever reached the monthly limit on your plan. If you have not, the tier below covers you.
The time saved went back into planning
A tool that saves you forty minutes and quietly absorbs them into a more elaborate unit has not saved you anything. This is the failure mode the trial hints at: the saving is real, but it never reaches your week.
Ask whether you finish planning earlier than you did last term. If you finish later, the tool is not the problem, the habit is.
Your board now pays for one that does the same job
A district licence landing mid-year is the single most common reason teachers are paying twice for the same saving. A second tool has to do something the first cannot to earn its price.
Name the one thing your subscription does that the board-funded tool cannot. If you cannot name it in a sentence, stop paying.
Why this decision comes due around week ten
The most rigorous trial in this category found a real saving and a real fade at the same time. The saving is why the tool is worth trying; the fade is why the subscription needs an audit rather than a renewal reminder.
The hard number
Lesson planning and preparation fell from 3 h 19 min to 2 h 30 min a week — about 49 minutes saved, a 24% reduction, with no measurable loss of quality.
The number that decides your renewal
Use of the tool declined over the ten-week trial, and fewer than one in five teachers felt it had meaningfully changed their planning.
Primary schools, one tool, ten weeks. It is evidence that early enthusiasm fades faster than the time saving does — not a forecast of your own usage.
Education Endowment Foundation / NFER — 464 teachers in 108 primary schools, published 6 October 2026The four-line ledger that settles it
Nobody remembers how often they used a tool. Three minutes on the first of each month, from your own generation history, replaces the memory with a number.
| Month | Generations you ran | Kept without a rewrite | Minutes saved, honestly | Price paid |
|---|---|---|---|---|
| — | — | — | — | — |
| — | — | — | — | — |
Four lines a month is the entire audit. If minutes saved sits under your break-even line for two months running, cancel — not because the tool is bad, but because for you it is not being used.
Test it, then step down before cancelling
Four weeks of ordinary teaching, no deliberate trials, then one subtraction. If you are not ready to end it outright, the cheaper tier usually delivers most of what cancelling would.
The four-week test
Do not schedule sessions to test the tool. A deliberate trial creates usage that does not survive it.
- 1Change nothing for four weeks. Do not schedule sessions to test it; a deliberate trial creates usage that does not survive the trial.
- 2Keep a tally of unprompted use — a tick on a sticky note, not a memory. Reaching for it while planning a real lesson is the only usage that counts.
- 3Note what you kept. A resource you would have made by hand anyway is a tick; one you rewrote from the first line is not.
- 4On day 28, subtract. Four ticks or fewer in a month means the subscription is buying a habit you no longer have, whatever the demo promised.
Step down before you cancel
- Drop a tier before you drop the tool. If you have never hit your monthly limit, the plan above the entry one is paying for headroom, and the cheaper tier does the same job.
- Drop to the free tier if your real volume is a couple of resources a month. That is what a free allowance is for, and it answers the question without spending anything.
- Cancel outright if you have not generated anything in three weeks. Downgrading a tool you do not open only postpones the decision by a month.
Five things to do before you click cancel
- Note your renewal date first. Cancelling two days after a renewal usually buys you a month you have already paid for.
- Export what you want to keep — PDF or DOCX of anything you would teach again — while your access is still live.
- Finish the unit you are in the middle of. Swapping tools mid-unit is the one moment where cancelling genuinely costs you time.
- Ask about downgrading before cancelling outright. Most tools would rather keep you on a smaller plan than lose you.
- Check whether a board or department licence already covers it. Double-paying is the most expensive version of this decision.
What this page is not
It is not a pitch dressed as advice. If the arithmetic says cancel, cancel — the free tier is enough to answer the question for most teachers, and the cheapest route to a finished lesson on this site is a lab that costs nothing at all.
The one case where we would push back on cancelling: if you are under the line because you never learned the tool, that is a different problem from the subscription being wrong. Read the cost of one lesson, measured before you decide the whole category does not work for you.
Questions teachers ask before cancelling
How do I know whether to cancel an AI teaching tool?
Put its price against the time it actually gives back. At
Do teachers really stop using AI tools after a few weeks?
In the 2026 Aila trial, use of the tool declined across the ten-week trial and fewer than one in five teachers felt it had meaningfully changed their planning, even though planning time did fall by about 49 minutes a week. That is one tool in primary schools, so treat it as a warning rather than a verdict: early enthusiasm fades, which is exactly why the decision should rest on your own usage record.
Should I cancel before or after my renewal date?
Before. Find the renewal date when you subscribe and put it in your calendar a week early. Cancelling after a charge has gone through rarely returns that month, so the timing of the decision is worth more than the argument about whether to make it.
Can I downgrade instead of cancelling?
Often, and it is usually the better first move. If you have never reached your monthly limit, the tier above your entry plan is buying capacity you do not use. Dropping a tier keeps your saved work and cuts the price, which is most of what cancelling would achieve.
What do I lose if I cancel?
Access to new generation and marking, not the work you have already made. Export the resources you would teach again as PDF or DOCX before the period ends, so that a later change in policy cannot put your own work behind a paywall.
Is it normal to cancel and come back later?
Yes, and it is rational. Teaching usage follows the school year: July and August are thin, September and exam season are heavy. Cancelling for the quiet months and returning when you are planning again is a sensible way to hold a subscription, and any tool that makes that hard is telling you something about how it sees you.
Nothing to cancel on the free tier
Run the four-week test before you ever pay. Generate one unit's worth, note what you kept, and decide with a number — no card, no renewal date, and no decision to unwind later.