A school has just explained how an AI tool may support a classroom activity. Teachers have revised the instructions, students have practised the boundaries and families have received a clear account of the approach.
Then the product changes.
The new feature may be useful. It may be irrelevant to the lesson. It may alter what students can do so substantially that the school needs to reconsider its approach.
Schools do not need to match the release cycle of AI products. They need to recognise which developments change an educational decision, respond to material concerns and give useful practices enough time to become dependable.
That requires attention, professional learning and organisational choices. Another tool can support parts of that work. It cannot decide how much change a school can responsibly take on.
“Keeping up” needs a better definition
The phrase can mean knowing that a development exists, understanding its implications, evaluating it, authorising a particular use or making that use routine.
These are different commitments.
A curriculum leader may need to know about a new capability without asking every teacher to adopt it. A teacher may need an explanation of how a change affects an assessment without learning the entire product. A school may need to stop a particular use while retaining another.
If every announcement becomes a demand for immediate classroom change, the organisation has accepted an agenda it did not choose.
A more useful definition of progress concerns the educational purpose. Can the school explain what students are learning, how assistance fits that aim and what evidence remains meaningful? Does a new development alter those answers?
The pace of announcements is context. The learning purpose should guide the response.
What current evidence contributes
New Zealand's Education Review Office published Ready or not: How are schools responding to Artificial Intelligence? on 30 July 2026. Its review describes uneven planning and support alongside widespread use.
Data were collected in Term 1 of 2026 from English-medium state and state-integrated schools with students from Year 5 upwards. The summary reports 859 teacher and 472 leader survey responses, alongside student and family responses, interviews, focus groups and eight case-study school visits. These are descriptive accounts and perceptions, not a causal test or a finding about every country's schools.
The Education Endowment Foundation's A School's Guide to Implementation, third edition published in April 2024, places attention on people's behaviour, context and a structured implementation process. It concerns educational implementation broadly, rather than a tested formula for the pace of AI adoption. The public overview was consulted here.
England's leadership support materials for AI in education, updated on 19 May 2026 for 2026–27, distinguish understanding opportunities, examining current use and planning within a wider digital strategy. Their existence is not evidence that every school has implemented them or that one jurisdiction's approach applies everywhere.
The argument below is an editorial interpretation: a school should manage the implications of change, rather than equate responsible adaptation with adopting more features.
Follow one change beyond the interface
Imagine a fictional geography department using an approved environment for a source-comparison activity.
Students read two prepared sources, form an initial interpretation and use a limited AI function to generate questions they should consider. They then select relevant questions, examine the sources and explain their own comparison.
Later, a new feature can generate a complete comparison directly from the material.
This does not automatically make the original activity worthless. It changes a condition the teachers need to examine.
Could students now bypass the reasoning the task was intended to develop? Are the existing instructions still intelligible? Can the limited function remain available without exposing students to the broader one? Does the department need to adapt the activity, restrict a use or choose a different approach?
These questions concern a fictional teaching scenario, not a verified product, actual school deployment or Learning Coach engagement.
The update itself is only the beginning of the possible change.
Count the human work required to respond
Suppose the department decides that its instructions need revision.
Someone must first understand what changed and whether it affects the intended use. Teachers then need to agree what students may do, prepare examples, adapt their explanation and consider how they will recognise the learning they intend to assess.
A colleague who missed the original discussion needs the same usable account. Students need to understand which instruction is current. Families may need clarification if the activity continues at home.
A small interface change can therefore generate several pieces of educational work.
That does not mean every update creates a major burden. Some require little more than a brief check. The point is to establish the work before promising that the school can absorb it immediately.
Ask the people affected what needs to be different in practice. Include preparation, explanation and follow-up, rather than counting only the time needed to discover a feature.
Separate a new capability from a necessary change
In the geography example, several responses remain possible.
The new feature might not be available in the approved environment. The department might retain the existing activity with clearer instructions. It might decide the additional capability offers no educational benefit for this task. Or it might investigate a different activity that makes constructive use of it.
None of those conclusions should follow solely from the fact that the feature is new.
Ask what problem the proposed change solves. Does it support an identified need? Does it alter something the school already relies on? What would become harder if teachers introduced it? What would be lost by waiting?
A school can be informed without making a commitment. Evaluation can reasonably end in a decision not to adopt.
For the wider work of translating an agreed position into everyday expectations, see our guide to moving from school AI policy to classroom practice. The question here comes earlier: which change deserves that implementation effort now?
Protect consistency without freezing practice
Students need understandable expectations. Teachers need enough continuity to see how an approach works.
If an activity is revised every time someone discovers a new option, it becomes difficult to establish what caused a misunderstanding. Was the difficulty in the learning task, the explanation, the tool or a change students had not understood?
Consistency provides a basis for examining those questions. It is not a reason to preserve an unsuitable practice.
In the fictional department, teachers could retain a clearly described approach while an identified colleague examines the implications of the new feature. If a material concern emerges, they should use the school's established route to reconsider or pause the affected use.
Some matters cannot wait for a convenient review meeting. A relevant safety concern, changed access condition or new requirement may demand prompt action. Appropriate specialists and the applicable local framework should guide that response.
The distinction is between deliberate continuity and inattention. Stability needs an active owner.
Make selective attention a shared responsibility
It is unrealistic to expect every teacher to follow every product announcement and independently interpret its consequences.
It is also risky to let all relevant knowledge sit with one enthusiastic colleague.
Schools can decide who watches which developments, how staff raise a concern and what warrants wider attention. The arrangement should fit the school's size and resources.
A curriculum question may need subject expertise. A change involving information handling may need the relevant technical or privacy owner. A new pupil-facing capability may need several perspectives before a decision is made.
The purpose is to route attention intelligently, not to create another committee for every notification.
Teachers should be able to bring an observation without also becoming responsible for solving the entire organisational problem. Someone needs to decide what happens next and communicate the outcome.
Where a small school lacks specialist capacity, shared support across an appropriate network or authority may be necessary. Local initiative should not disguise a need for resources beyond the school.
Give implementation time a real place in the workload
A request to “try the new version” may sound modest. It can still require a teacher to reconstruct materials, learn an unfamiliar sequence and answer new student questions.
That work competes with existing responsibilities.
Before adding a change, leaders should ask what time it requires and what can realistically receive less attention. If nothing can move, the proposed timing may need to change.
A bounded investigation can be more useful than asking the whole staff to experiment independently. But it still needs an appropriate purpose, scope, support and review.
Avoid making the most confident users the permanent unpaid support system for everyone else. Helping colleagues, preparing examples and maintaining guidance are contributions that need recognition and capacity.
Technology may assist with drafting or organising these materials. People remain responsible for checking whether they explain the actual decision accurately.
Retain enough memory to avoid starting again
When a school changes an approach, it should preserve a concise account of what changed and why.
For the geography department, a useful record might explain the activity's learning purpose, the permitted assistance, the reason for the latest revision and the unresolved question that will trigger another review.
A new colleague can then understand the reasoning, rather than inherit an unexplained rule.
This need not become a large archive. The record should be proportionate and easy to find. Outdated instructions should not remain indistinguishable from the current version.
Memory also helps a school recognise when it is revisiting an old question. A newly marketed feature may raise a distinction the department has already considered. The existing reasoning may still apply—or a material difference may justify reopening it.
The aim is cumulative learning, not constant reinvention.
Ask a different question at the next school event
A school AI conference can unintentionally intensify the pressure to keep up. Each session introduces another possibility; teachers return with more ideas than the organisation can examine.
An alternative event brief could ask participants to follow one fictional change through the school.
Who notices it? Who understands its implications? Which materials would need revision? Who explains the decision to students? What remains stable? Which concern would require a faster response?
A keynote can establish that perspective. A working session can apply it to appropriately prepared examples. Neither should imply that attendance establishes readiness for every tool.
Our guide to choosing an AI in education keynote speaker helps organisers connect the speaker's contribution to an educational question.
The useful event output might be a clearer priority or an unresolved decision with an owner. It need not be a commitment to adopt something new.
Choose the change you can explain and support
Sources were checked on 28 October 2026. This article uses a New Zealand review, a general implementation guide and England-specific support material. They address different contexts and do not establish an optimal rate of change. The full EEF report and the DfE's attached leadership slides were not consulted.
The fictional geography example is a way to examine the work involved, not evidence of an intervention's effectiveness.
Learning Coach's position is that a school should remain alert while choosing carefully. Product awareness, educational judgment and dependable classroom practice require different kinds of attention.
Which change would improve learning enough to justify the human work of making it coherent—and which can wait?
Bring that question to your next school leadership conference or professional learning event. Explore Learning Coach's international keynotes and request a tailored proposal.
References
- Education Review Office, New Zealand. Ready or not: How are schools responding to Artificial Intelligence? Summary Report. 30 July 2026. Summary findings and methods consulted; Term 1 data.
- Education Endowment Foundation. A School's Guide to Implementation. Third edition, 24 April 2024. Public overview consulted.
- Department for Education, England. Using AI in education: support for school and college leaders. Updated 19 May 2026 for 2026–27. Public resource description, not a universal implementation requirement.
