What Is Your AI Problem?

AI Is Supposed to Save Teachers Time. Does It?

By Dr. Jamiylah Jones, CEO of Creative Transformations · Nov 18, 2026 · 8 min read

Consider a teacher getting ready for the next day. There is a lesson to plan, feedback to finish, and a family message that still needs a response. The teacher asks AI to create an activity, a few questions, and some supports for students who may need them. Within minutes, the materials appear. On a busy afternoon, that can feel like a real help.

Then comes the review. One question does not quite fit the learning goal. A suggested support gives away the answer instead of helping students work through it. The activity looks complete, but it does not give the teacher a clear way to see what students understand on their own. The teacher makes changes, reads through everything again, and decides what is ready to use.

AI helped with the first draft. But how much time did it save by the time the lesson was actually ready? That is harder to see in a quick demonstration. The finished materials are right there on the screen, while the checking, adjusting, and decision-making may happen during a planning period or later that evening. Those parts of the work still count.

This is a shift schools need to pay attention to. Some of the time teachers once spent creating materials may now be spent reviewing what AI creates. Sometimes that is a useful trade. Other times, the teacher may wonder whether it would have been easier to begin with a familiar resource. Both experiences deserve a place in the conversation about AI and workload.

A recent report from New Zealand’s Education Review Office described some of this complexity. Teachers reported ways AI made their jobs easier, but the report also identified additional work in secondary schools to check the authenticity of student work. Those findings come from the schools surveyed, so they do not tell the story of every school. They do remind us to look beyond how quickly AI produces something and ask what happens next.

Why It Matters

Teachers have always checked resources, adjusted lessons, and made decisions about what students need. AI does not remove those responsibilities. It can, however, produce much more material in a short time, and a polished resource can look ready before it really is. The teacher still has to decide whether the work is accurate, whether it fits the lesson, and whether students will be doing the thinking the task is meant to develop.

When something gets missed, the work can show up later. Unclear directions may need to be explained again during class. A support that does not fit may need to be replaced while the lesson is underway. Feedback that sounds helpful but does not match a student’s work can leave the student confused and the family asking questions. What looked like a time-saving step during preparation can create more work during the day.

There is also a difference between learning a new process and getting stuck with a process that does not work well. A teacher may need some time to become comfortable with an approach that eventually helps. But if the same problems keep appearing, more practice may not be the answer. The task may need to be narrowed, the tool may not be a good fit, or the school may need to change what it is asking teachers to do.

The reason for saving time matters, too. Teachers need space to respond to students, think through a lesson, work with colleagues, and finish the day without carrying so much home. If AI makes it easier to produce documents and the school responds by asking for more documents, there may be very little relief. The documents get produced faster, but the amount of work keeps growing.

Practical Applications

Schools can start small. At the next grade-level or department meeting, teachers could bring one AI-generated resource and the version they actually used. What was useful? What needed to change? Did anything create extra work once students began the activity? Looking at a real example can tell a team much more than asking, “Is AI saving time?” It can also fit into a conversation about instruction that is already happening, rather than becoming another task to document.

For classroom teachers, this may lead to a more focused way of using AI. A teacher who already has a suitable activity may need only clearer directions or a way for students to organize their responses. Asking for several new versions of the whole lesson creates more material to review. A smaller request can make the review more manageable because the teacher knows exactly what the output is supposed to help with.

For example, students may have ideas but struggle to put an explanation together. AI could help draft a planning structure with space for an idea, supporting information, and an explanation of how the two connect. The teacher can check whether it helps students organize their thinking without doing the thinking for them. If that support works, the teacher has a useful starting point for a similar task later. If it does not, an existing resource or a teacher-created example may be the better choice.

Instructional coaches and team leaders can look for problems that teachers are solving separately. If several teachers keep revising questions that do not match the learning goal, it may help to agree on what a useful question should do and develop a reviewed example together. If generic materials take too long to adapt, the team may need a shared starting point that fits the curriculum. After a few uses, the team can return to the example and discuss whether it helped. The aim is to avoid repeating the same work in every classroom while still leaving room for teachers to make decisions for their students.

Special educators and multilingual learner staff can help teams think through supports before unsuitable adaptations become part of the routine. During an existing planning or student-support meeting, they can identify what students need help accessing and what the task still needs to ask them to do. Shorter directions may help a student get started. Removing the need to explain an answer changes the learning itself. These conversations should also protect documented accommodations and clarify what information can be entered into approved AI tools. A process that saves general planning time but leaves specialists correcting materials afterward may simply be moving the workload.

Counselors and other student-support staff may notice this when AI is used for summaries or family communications. A draft can sound clear and still miss the meaning of what a student or family shared. The team can decide which routine tasks are suitable for an approved tool, what must be checked, and when a note written directly by a staff member is more appropriate. Confidentiality, context, and uncertainty all matter here. If reviewing an AI-generated summary takes more effort than writing an accurate note, staff should be able to say so and choose another approach.

Principals can help by looking at the requirements around the work. At a leadership meeting, they could choose one recurring document and ask who reads it, what decision it helps with, and whether the same information is already recorded somewhere else. If it duplicates another requirement, removing it may help more than making it faster to produce. Principals can also make clear that checking AI-supported work is part of the task and requires time. Teachers should not feel that raising a concern about extra work will be taken as resistance to change.

District and technology leaders can carry this into decisions about tools. A demonstration may show what a product can generate, but a useful pilot needs to show what happens when educators try it with their actual work. How much needs to be revised? Does the tool fit the curriculum? Can staff use it without entering the same information in several places? Does it support the learners it is meant to serve? Those questions can help leaders decide whether a tool is worth continuing, whether it should be used for a smaller set of tasks, or whether another approach would serve the school better.

What Schools Can Take Away

The first takeaway is to look at the whole task. Producing a draft is one step. Checking it, making changes, using it, and responding to what happens afterward are part of the work as well. Schools do not need a complicated tracking system to begin understanding this. A few honest conversations around real examples can show where AI is helping and where more support is needed.

The second is that teachers should not have to work this out alone. Teams can share useful examples, coaches can help connect materials to learning goals, specialists can identify access needs, and leaders can remove work that no longer serves a clear purpose. Each group has a different contribution to make. When those contributions come together, schools have a better chance of reducing repeated work rather than passing it from one person to another.

The third is to be clear about the benefit. AI might save time, improve a resource, or help an educator complete something that was previously difficult to manage. Each can be worthwhile, but they are not the same outcome. A better resource may still take time to prepare. A faster process may still need to change if the result does not help students. Schools need room to recognize what is working and to stop or adjust what is not.

When a school says AI is saving teachers time, it is worth asking where that time went. Is there more room to listen to a student, respond thoughtfully to a family, or prepare for the next day without working late? Has the work become easier to manage, or has it simply changed form? Those questions bring the conversation back to why workload relief matters in the first place.

This is one of many hidden shifts AI is introducing into education. I write these newsletters to help educators and school leaders see them early and respond intentionally. If that matters to you, stay with the series.

Source: Education Review Office, “Ready or Not? How Are Schools Responding to Artificial Intelligence?”

Keep reading each week

New editions publish weekly. Subscribe on LinkedIn or read the full archive here.

Previous edition

The Score Went Up. Did the Learning?

Next step

Want this thinking applied to your school?

Tell us the AI problem you are facing and we will show you what it looks like solved.