A school has an approved AI service, several enthusiastic users and a growing list of unresolved questions. Another application promises to connect the pieces. An external adviser offers to help the school decide what it actually needs.
Either purchase could add value. Either could add another layer of work.
A school may need an independent AI implementation partner when a specific gap in expertise or capacity prevents it from making and carrying through a sound educational decision. The case for outside help depends on the missing work, the adviser’s competence and interests, and what the school will be able to do afterwards.
Independence is a condition to examine, rather than a title to accept. A partner should be able to recommend a different product, a smaller intervention or no additional purchase—and explain the evidence behind that recommendation.
Identify the work before the provider
“Help us with AI” is too broad to commission usefully.
A school might need technical integration, staff development, help examining a proposed classroom use, support with buying or a research evaluation. These require different skills and produce different outputs.
A supplier may be well placed to explain its own product. A school-network colleague may help compare teaching experiences. A specialist adviser may connect evidence, local requirements and implementation choices. One organisation may offer several services, but their boundaries still matter.
Start with the decision the school cannot currently make or the work it cannot complete. Perhaps nobody has examined whether a proposed workflow fits teachers’ available time. Perhaps the school understands its requirements but needs help comparing options. Perhaps it needs additional capacity to organise a carefully bounded implementation.
If the gap is simply an unanswered product question, paying for a broad consultancy programme may be disproportionate.
What implementation guidance can tell us
The Education Endowment Foundation’s third-edition A School’s Guide to Implementation, published on 24 April 2024, treats implementation as more than introducing an idea. Its public overview connects behaviours, context and a structured process; its accompanying behaviour guidance emphasises engaging people, establishing shared direction and reflecting on implementation.
That is general school-improvement guidance. It does not establish that an external AI adviser is better than an internal team.
A 2026 invited commentary by David Hollow and Verna Lalbeharie in the British Journal of Educational Technology adds a useful distinction. Drawing on EdTech Hub’s implementation-research experience, principally in low- and middle-income countries, the authors separate implementation from research about implementation. They also discuss the tension between researchers remaining independent and working closely enough with implementers to be useful.
This is a methodological and practice-based argument, not a trial of school consultants. Its relevance here is to clarify roles before treating a partner’s involvement as independent evidence.
The practical proposals below apply that caution to commissioning external support. They are not a validated selection instrument.
Ask what “independent” means in practice
An adviser can be legally separate from a software company while depending on that company for referrals, training access or future work.
Ask about relevant relationships. Does the organisation resell licences? Receive referral payments? Hold a formal supplier partnership? Obtain funding from a provider whose product it recommends? Will its fee change if the school chooses a particular option?
These questions do not establish wrongdoing. They help the school understand how an incentive could affect a recommendation.
Expertise in one platform can be valuable. It should be described accurately as platform-specific expertise, rather than presented as a neutral comparison of the whole market. Likewise, a declared commercial relationship does not automatically invalidate the advice. Leaders need to judge whether it is compatible with the work being commissioned.
Independence also concerns the mandate. If the school has already decided to adopt a product, it should not describe a project to support that adoption as an open investigation of whether to buy it.
Require a meaningful comparison
A recommendation becomes easier to assess when the alternatives are visible.
Consider a fictional geography department preparing pupils to design a fieldwork questionnaire. Teachers want pupils to recognise leading questions, choose relevant information and explain why their questions suit the investigation.
An existing AI service can generate questionnaires quickly. A new platform offers a more organised workflow. The school wonders whether it needs the additional software.
An implementation adviser could compare three routes: a shared bank of teacher-prepared examples, a teacher-reviewed use of the existing service, and the proposed platform. Each route would be examined against the same educational purpose and practical constraints.
Which route gives pupils opportunities to identify and revise a leading question? What preparation and checking does it require? Can teachers see the reasoning behind pupils’ choices? What access arrangements are needed?
The point is not that the simplest option always wins. It is that the comparison should leave that possibility open.
This is an illustrative commissioning case, not an account of a school deployment or a Learning Coach client.
Commission a first answer before a large programme
Where the need is uncertain, a limited discovery stage may be useful.
In the geography example, its output might be a short account of the learning problem, the available approaches, unresolved requirements and the next decision. It could include a rehearsal with fictional materials and a clear explanation of what that rehearsal cannot establish.
That is more concrete than a promise to “accelerate transformation.” It gives the school something to examine before committing to further work.
Agree who needs to contribute. An adviser cannot understand a classroom process solely through a leadership presentation. Teachers, relevant support staff and accountable institutional owners may hold different parts of the picture.
Also agree the school’s contribution. If staff have no time to supply examples or examine findings, an external appointment will not make that constraint disappear.
A discovery stage should be able to conclude that further consultancy is unnecessary.
Check competence separately from commercial independence
An adviser with no supplier ties can still give weak advice.
Ask for evidence that the proposed team understands the relevant age group, subject, institutional context and kind of implementation. Request an explanation of how they distinguish a useful demonstration from a claim about learning.
A practical discussion can be revealing. Present a fictional request with incomplete information and ask what they would need to know before recommending an approach.
For the geography case, a thoughtful response would examine the fieldwork objective and the work pupils must do. It would also consider teacher review, access and the school’s approval arrangements. A response that immediately selects an application leaves important questions unanswered.
Check experience claims and references where they are relevant and available. Do not infer that a confident presentation demonstrates technical assurance, research competence or knowledge of every school’s obligations.
The school needs the expertise required for its actual problem.
Keep implementation, procurement and evaluation distinct
Outside support comes in several forms.
England’s Department for Education provides Get help buying for schools, a free service for schools and trusts in England. Its current service page describes help running procurements and finding approved buying options. That illustrates an available support route, not a universal entitlement or an AI pedagogy service.
A procurement adviser helps with buying. A technical provider helps configure and maintain systems. A professional-learning partner helps staff develop relevant capability. A researcher investigates a defined question through appropriate methods.
These roles can collaborate. They should not be treated as interchangeable.
A team that helps implement a tool may be able to describe its work accurately. That does not make its account an independent evaluation of the tool’s effectiveness. If the school needs an external judgment, it should specify how the reviewer’s relationships, access to evidence and reporting arrangements will support that purpose.
Use the school’s own procedures and qualified specialists for contractual, data, safeguarding and other formal decisions.
Make the deliverables useful after the engagement
A presentation and a collection of links may help people begin. They may not resolve the work for which the partner was hired.
Ask what usable outputs the engagement will leave. Depending on the remit, these could include a documented comparison, a clarified workflow, a staff rehearsal, a list of unresolved decisions and an agreed route for maintaining the work.
In the geography example, teachers might retain a process for examining a generated questionnaire against the learning objective. They should understand how to use that process with a later task, rather than wait for the adviser to review every example.
Specify how the school can access and maintain the materials it needs. Clarify any continuing service, update requirements or reliance on the partner’s own systems before assuming that the handover is complete.
Building school capability does not require every teacher to become a technical expert. It requires enough understanding and appropriate support for the school to continue the essential work.
Leave the school’s decisions with the school
An external partner can make responsibilities seem clearer while quietly becoming the only person who understands them.
Name the institutional owner for each consequential decision. The adviser may help assemble information and explain alternatives, while the appropriate school or trust leaders decide what to approve.
Subject teams still need to judge curriculum fit. Relevant specialists still need to examine matters within their remit. The partner should identify the limits of its competence and refer questions accordingly.
Our guide to turning AI policy into classroom practice develops those wider connections between principles, responsibilities and everyday routines.
The school should be able to explain a decision in its own words. “Our consultant recommended it” is an incomplete account of why it is educationally appropriate.
Review contribution without confusing activity and results
Meetings held, staff trained and resources produced describe what a partner did. Leaders also need to ask whether that work addressed the original gap.
Can the school now compare the relevant options? Has an unresolved workflow become understandable and feasible? Do staff know which decisions remain open and who owns them? What continuing support is genuinely necessary?
Where the engagement concerns a learning intervention, the claim about learning needs its own evidence. A well-organised implementation does not establish that a pupil learned more.
Equally, an adviser who helps a school avoid an unsuitable purchase may have made a useful contribution without increasing adoption. Evaluation should allow for that outcome.
Compare the contribution with the burden it created, including staff preparation, meetings and maintenance. The review can support continuation, a narrower remit, a different form of help or an end to the engagement.
Use a leadership event to clarify the brief
A school-network leadership event can examine a fictional offer of implementation support.
Give participants the same educational problem and several possible remits. Ask which missing work each remit addresses, which interests need disclosure, what the school must contribute and what output would justify the next step.
A keynote can establish the distinction between acquiring tools and building the capacity to judge and use them. A working session can develop a more precise brief. Neither is a substitute for product assurance, procurement or research evaluation.
Our guide to choosing an AI in education keynote speaker helps organisers connect that conversation with their audience’s decisions.
This article draws on public implementation guidance, selected sections of an invited research commentary and an official service description. It does not establish a market trend, compare named consultancy providers or show that external partners improve school outcomes. The proposals require local adaptation.
An independent AI implementation partner may be useful. The strongest case is a defined piece of work the school needs help to complete, with advice it can question and capability it can retain.
Planning a school leadership event about the human work behind AI implementation? Explore Learning Coach’s international keynotes and request a tailored proposal.
References
Education Endowment Foundation, A School’s Guide to Implementation. Third edition, 24 April 2024; official overview and accompanying behaviour guidance.
Hollow, D., and Lalbeharie, V. The role of EdTech Implementation Research in connecting evidence and practice: Insights for a high-impact future. First published 19 April 2026. Invited commentary, not a comparative trial of implementation services.
Department for Education, Get help buying for schools: About our service. Undated current service description; scope restricted to the stated England service. Sources consulted 29 October 2026.
