Michael Harvey is a New Zealand teacher who writes about education technology, AI, and how to design learning that keeps teachers and students in control.
2.9K followersAI in educationBased on 11 posts · latest Oct 2026
Michael Harvey is useful for testing how education AI tools maintain teacher and student agency.
Best for
Education AI companies seeking a practitioner to test if features support meaningful learning.
Not for
Campaign requirements for platform endorsements, mass reach, or evidence of classroom results.
Key takeaways
AI should widen choices for teachers and students, not make them.
Pedagogy must guide technology choices, not the brand logo.
Innovation requires testing unfinished ideas before they are fully formed.
Check first: Confirm audience demographics, such as roles, countries, and seniority, and request recent results.
Main ideas
AI should widen choices for teachers and students, not make them · Neither technology companies nor education companies should own the design of learning · Pedagogy first, whichever technology ecosystem it comes from
Michael Harvey is a New Zealand teacher who writes about education technology, AI and the choices that shape learning.
They argue that education is not just about finishing tasks efficiently: the thinking involved matters, and teachers and students should retain the power to decide when AI helps.
They also argue that pedagogy matters more than the technology brand.
Follow for perspectives on learning agency, classroom AI and educator-led experimentation.
Michael Harvey connects technology changes to practical questions about who gets to make decisions in learning.
In a post on Google’s move from Gems to Skills, they consider how AI might use teaching practices such as Socratic questioning and formative feedback, then ask whether teachers and students will remain able to choose, question or reject that support.
The point is not simply to make AI more capable: a tool should expand people’s choices, not make learning decisions for them.
They say pedagogy—not the Google or Microsoft logo—should guide technology choices, and see value in connecting education communities rather than treating them as competing camps.
Their reflections on professional learning emphasize the same openness: educators should challenge assumptions, listen to different perspectives and test ideas before they are fully formed.
These posts are useful to teachers and education leaders weighing how AI and other tools can support meaningful learning without displacing human judgment.
Michael Harvey argues that AI and education technology should widen what teachers and students can choose to do, not decide for them.
Read more
They put pedagogy ahead of tools and vendors as the driver of good learning, and describe innovation as testing half-formed ideas with other educators rather than waiting for a breakthrough.
AI should widen choices for teachers and students, not make them
Harvey sees Google's shift from Gems to Skills as more than a product change.
Read more
Skills could let teachers encode practices such as Socratic questioning, misconception diagnosis, source evaluation and formative feedback as reusable capabilities—moving beyond prompt engineering to what they call portable pedagogical design.
The question is who decides which capability gets invoked.
If platforms increasingly choose when support appears and what a student needs, judgement shifts from teachers and students to platforms.
Harvey's test: do teachers retain the call on why, when and how AI supports learning, and can students seek help, question it, reject it or work through a problem first?
Harvey argues that education is not productivity.
A teacher may deliberately withhold a scaffold or require a prediction before explaining, while a student may struggle with an idea before seeking AI support.
They see these choices as expressions of teacher and student agency, not inefficiencies to engineer away.
Neither technology companies nor education companies should own the design of learning
Harvey credits Becky Keene with a useful technical distinction: a Gem is somewhere you go, whereas a Skill is something AI uses when that capability is needed.
Read more
They also report Keene's suggestion that large AI companies will supply the intelligence while education-specific companies supply learning design.
Harvey finds the distinction useful but rejects the framing: learning design should belong to neither kind of company.
Teachers and students should retain meaningful agency over it.
Harvey closes with the question at the heart of the argument: how do we encode good pedagogy into AI while ensuring teachers and students remain the agents of learning?
Pedagogy first, whichever technology ecosystem it comes from
Selected as New Zealand's Microsoft Elevate Education Fellow for 2026-28 in the same week as becoming a Google Innovator, Harvey rejects the idea of choosing sides.
Read more
Technology matters only when it helps design better learning, strengthen relationships, give students greater agency and support teachers. They want to connect these ecosystems rather than treat them as competing camps.
Harvey writes that good education has never been about the logo on the tool, and jokes that after three days immersed in Google, they still prefer Microsoft Teams for administration.
Innovation means testing unfinished ideas, not waiting for a eureka moment
Reflecting on three days at the Google Innovator Academy in Melbourne, Harvey says the biggest takeaway was not a Google tool but people, pedagogy, curiosity and permission to test ideas that are not fully formed.
Read more
Their account of innovation is deliberately unglamorous: try something and see what happens. The graduation post frames the end of the programme as the start of the work, with coaches and fellow educators turning rough ideas into more purposeful projects expected to change, grow and sometimes be completely reshaped over the following year.
Harvey describes innovation as less 'eureka' and more 'well... let's suck it and see', and says graduation does not feel like the end of the programme.
Learning from teachers who see the same problems differently
Invited to review sessions for ISTELive27, Harvey highlights what they want to learn from: ideas that challenge assumptions, celebrate student voice and make teaching and learning meaningful in an increasingly AI-influenced world.
Read more
They say some of the most valuable professional learning comes from conversations with teachers who see familiar challenges through completely different lenses, and note the value of contributing to a global community from a New Zealand classroom.
Harvey is interested in sessions that question assumptions and put student voice and meaningful learning first, rather than those organised around technology adoption.
This is a direct, concrete statement of Michael Harvey’s position on AI in education—and a useful way to understand what they mean by teacher and student agency.
Read more
Michael Harvey uses Google’s move from Gems to Skills to make a broader point about agency in AI-supported learning.
Skills could let teachers build practices such as Socratic questioning and formative feedback into AI, but Harvey argues that neither technology nor education companies should own the design of learning.
Teachers should decide when and why AI is used; students should be able to question its suggestions, reject them, or choose to work through a problem first.
In education, the thinking involved in a task can matter more than completing it quickly.
The design principle is clear: AI should give teachers and learners more choices, not make those choices for them.
For brands planning a campaign
For an education-technology campaign, this argument favours showing how a product gives teachers and students meaningful control over AI support. A message focused only on speed or automation would sit uneasily with Harvey’s emphasis on learning through thought and choice.
The post makes Harvey’s cross-platform outlook explicit: the education purpose matters more than the logo.
Read more
After being selected as New Zealand’s Microsoft Elevate Education Fellow for 2026–28, Michael Harvey reflects on also becoming a Google Innovator.
Rather than treating the two affiliations as competing loyalties, they argue that pedagogy—not the brand of a tool—should drive decisions.
Technology matters when it helps design better learning, strengthen relationships, give students agency, and support teachers.
Harvey even notes, with humour, a continuing preference for Microsoft Teams for administration after spending time immersed in Google tools.
For brands planning a campaign
A campaign centred on learning outcomes or a genuinely cross-platform approach fits this stated outlook better than one asking Harvey to take sides between vendors.
This reflection shows Harvey’s process-oriented view of innovation: experiment, learn, and keep the focus on teaching and learning rather than the technology itself.
Read more
Reflecting on three days at the Google Innovator Academy in Melbourne, Harvey says the main takeaway was not a tool but the value of people, pedagogy, curiosity, and testing ideas before they are fully formed. Their description of innovation is less about a sudden breakthrough and more about trying something and seeing what happens.
For brands planning a campaign
The post suggests a natural fit for a workshop or process-led series that shows educators testing an idea, rather than a polished demonstration that presents the outcome as inevitable.
This offers a concise view of the educational ideas Harvey wants to hear more about, as well as the importance they place on learning across a global teacher community.
Read more
Harvey announces an invitation to review sessions for ISTELive27 and describes the value of learning from teachers who see familiar challenges differently. The ideas they are especially interested in include those that challenge assumptions, celebrate student voice, and explore meaningful teaching and learning in an AI-influenced world.
For brands planning a campaign
An education campaign or event contribution is more likely to resonate if it puts student voice and meaningful learning at the centre, rather than treating technology adoption as the goal.
Educator-led experimentation and professional learning3
AI that preserves teacher and student agency1
Pedagogy before the technology brand1
AI that preserves teacher and student agency
Harvey argues that AI in education should support learning without deciding for teachers or students.
Read more
In a post about the shift from Google Gems to Skills, they describe reusable capabilities such as Socratic questioning, diagnosing misconceptions, evaluating sources, and giving formative feedback. The post warns that automatically choosing which capability to use could transfer decisions from people to platforms.
Makes the case through a product change, specific classroom examples, and a closing design question.
For brands planning a campaign
A brand developing education AI could assess Harvey for a design discussion or campaign about agency-preserving features. The post offers concrete questions for that work: who chooses when support is used, can a student question or reject it, and can a teacher deliberately withhold a scaffold?
Harvey presents tools as secondary to learning goals, relationships, and student agency.
Read more
In a post describing selection as both a Microsoft Elevate Fellow and a Google Innovator, they reject treating the two technology ecosystems as opposing camps and say technology matters when it helps educators design better learning.
For brands planning a campaign
A cross-platform education technology campaign could explore how a tool supports teaching rather than asking Harvey to endorse a brand rivalry. The post gives a clear basis for discussing interoperability and pedagogical purpose.
Educator-led experimentation and professional learning
Harvey describes innovation as testing ideas that are not yet fully formed, learning from peers, and improving work through questions and feedback.
Read more
The MEL26 reflection emphasizes design sprints, coaching, and community; the graduation post frames the programme as the start of continued work. A separate ISTELive27 post values learning from teachers with different perspectives and highlights student voice and meaningful teaching in an AI-influenced world.
Uses first-person reflections on programmes and educator communities rather than step-by-step tool tutorials.
For brands planning a campaign
A professional-learning or education-innovation brief could focus on how educators test and refine ideas together, with student voice as a consideration. These posts support discussion of the process and values; they do not provide evidence of a specific classroom intervention or its measured results.
Michael Harvey writes about AI and education from a classroom teacher's perspective.
Read more
A useful entry point is the post on Google's shift from Gems to Skills, which poses a question that runs through their work: who decides when AI support appears, and can a teacher or student refuse it?
Two Melbourne Innovator Academy posts show their working method—design sprints, coaching and ideas tested before they are finished.
The ISTELive27 post shows what they look for in other people's sessions: challenges to assumptions, student voice and teaching that stays meaningful as AI spreads.
The tags cover AI in education, pedagogy, educator development, and student and teacher agency.
Similar creators approach the same territory through edtech venture building, product design and teacher preparation.
AI in educationPedagogyTeacher agencyStudent agencyEducation innovationTeacher professional developmentEducation technologyAI and pedagogyClassroom teachingLinkedIn postsFirst-person reflectionK-12 educatorsEdtech product teams
Both write about how people learn and how adults in the system can support that learning, translating research into practical design questions for programs and classrooms.
Read more
Harper's posts centre on early childhood and PK-3 teacher preparation in Texas, including a state scholarship and apprenticeship grant and a Learning Policy Institute framework for prep program design. Harvey writes from inside a K-12 classroom about AI, agency and pedagogy, rather than preparation policy or workforce funding.
A former teacher who designs education products, Lapis works on the same bridge Harvey describes between pedagogical intent and what a product lets a learner do, with student collaboration and equitable access as the goal.
Read more
Lapis's posts cover building and iterating an app, Studywavve, including onboarding, resource filtering, scheduling and messaging, as well as an earlier career-change post. Harvey argues about who controls AI-mediated learning and how educators should test new ideas, without describing a product build.
Both see education innovation as moving through people and ecosystems, not just tools.
Read more
They draw on conference experience to argue about where education is heading and what founders and educators should watch.
Greenwald writes from the edtech venture side, covering SXSW EDU, the GESAwards global startup program and the difficulty of selling into school systems. Harvey writes from the classroom side about pedagogy, teacher and student agency, and how to design AI support.
The only published audience figure is Michael Harvey’s LinkedIn following: about 2,900 followers, recorded on 8 October 2026.
Read more
For an edtech company building AI tools that preserve teacher and student agency, that’s a small, focused professional audience rather than a mass one.
The posts argue about pedagogy, teacher and student agency, and how AI should and should not make decisions for learners.
The communities they name—ISTE, the Google Innovator programme and Microsoft Elevate—are educator communities, making this content a fit for campaigns where credibility with teachers and education leaders matters more than raw reach.
The posts don’t identify followers’ countries, roles, institution types or seniority, so the follower count doesn’t verify that they are teachers, school leaders or AI procurement decision-makers.
The content itself consistently addresses people who work in or around schools and are thinking about AI adoption.
A post on Google’s shift from Gems to Skills argues that teachers and students must retain agency over why, when and how AI supports learning, and states plainly that education’s purpose is not task efficiency.
A post marking the Microsoft Elevate fellowship and Google Innovator badge argues that pedagogy, not the tool brand, drives good learning.
The posts provide no impressions, comments, reshares or click data.
Ask for recent results on comparable topics, depth and formats—for example, a text post on AI and pedagogy versus a programme-reflection post—and for LinkedIn analytics, audience survey results or another documented source showing where followers are based and what roles they hold.
If the campaign needs to reach teachers in a specific country or institution type, or people with budget authority, request evidence for those segments.
Michael Harvey's five LinkedIn posts (1–7 October 2026) describe vendor relationships as educator-community programmes, not sponsorships, advertising or paid collaborations: a Google Innovator cohort place at the Melbourne academy (#MEL26), a Microsoft Elevate Education Fellow place for 2026–28, and an invited session-reviewer role for ISTELive27.
Read more
They frame these as professional development and community, and make clear that the tool brand is not the deciding factor — the pedagogy is.
Their focus on teacher and student agency, encoding good pedagogy into AI, and questioning which skills an AI invokes and when could suit a brand seeking critique or context from a practising teacher, rather than a straight promotional slot.
Before planning outreach, ask for their sponsorship policy, media kit, pricing, booking route and any available package or deliverables.
A practical fit is an education AI company that wants a practising teacher to test whether its features leave real decisions with teachers and students.
Read more
Michael Harvey's post on Google's move from Gems to Skills sets out the test a product would have to pass: a Gem is somewhere you go, while a Skill is a capability the AI invokes when needed.
The risk is that the platform quietly decides which support a learner gets and when.
The post names the classroom moves worth building in — Socratic questioning, diagnosing misconceptions, source evaluation, formative feedback — then makes the argument that matters for a design brief: teachers should decide why, when and how AI supports learning; students should be able to question, reject or postpone what AI offers; and a teacher deliberately withholding a scaffold is not an inefficiency to be designed away.
Harvey also brings credibility across the two ecosystems an education AI buyer usually has to consider.
A post announcing both a Microsoft Elevate Education Fellow place for 2026–28 and a Google Innovator place argues that pedagogy, not the tool brand, determines whether technology is worth using, and that technology matters when it helps design better learning, strengthen relationships and give students more agency.
That makes Harvey a reasonable person to review a feature set without the conversation collapsing into a vendor comparison.
Two constraints shape how to use this.
First, the only measured audience figure is a LinkedIn following of about 2,900, observed on 8 October 2026, with no country, role or institution breakdown.
Treat this as a credibility-led contribution rather than a reach play.
Second, no rate card, deliverable list or booking route appears in the material reviewed; scope and terms will need to come from a direct conversation.
When to choose a specialist
A campaign seeking a straight platform endorsement, a mass-reach launch push or evidence of classroom results does not fit what Harvey's posts say.
Read more
Harvey's posts argue against the framing such campaigns usually need.
The Microsoft Elevate and Google Innovator post refuses to treat the two technology ecosystems as competing camps and says good education has never been about the logo on the tool.
Asking Harvey to front a “switch from X to Y” or “our platform beats theirs” message would put the campaign directly against a position stated plainly in their own writing.
A message built on speed or automation also sits badly.
The Gems-to-Skills post says education has a fundamentally different purpose from productivity, and that the thinking required to finish a task is often the point: a teacher may withhold a scaffold, and a student may choose to struggle before asking for help.
A campaign promising to remove that effort would contradict the argument that makes Harvey worth hiring.
Mass reach is a poor expectation on the numbers available: roughly 2,900 LinkedIn followers, with no audience breakdown and no impressions, comment or reshare figures for any post.
A campaign needing measured classroom impact or product-testing results will also need evidence beyond what is available: the posts describe programmes, positions and reflections, not a classroom intervention with outcomes.
None of this rules out a paid collaboration, but no post documents one.
A promotional-slot campaign would have to be built from scratch rather than from a published format.
A worked campaign example
For an education AI company developing features that protect teacher and student agency, a design-review series follows most directly from what Harvey already does.
The core asset is an audit checklist drawn from Harvey's own questions: who decides which AI support is offered and when; can a student question, reject or postpone it; can a teacher deliberately withhold a scaffold; and does the feature leave the thinking that was the point of the task in the learner's hands?
Harvey then applies the checklist to one real prototype feature in a recorded working session with the product team.
The session is raw material, not the campaign.
From that session, publish three pieces: a plain-language teardown of the feature against the checklist on the company's educator-facing channel, cross-posted to Harvey's LinkedIn with the checklist attached; a short “what we changed and why” note from the product team responding to one point Harvey raised; and the checklist as a downloadable resource other educators can use to assess any AI tool. A follow-up dispatch later in the term can report whether the change held up in use.
Why this shape works: Harvey's strongest post takes a product change, names specific teaching practices and closes with a design question rather than a verdict.
Their Innovator Academy reflection treats innovation as testing an idea before it is fully formed rather than as a sudden breakthrough.
Their ISTELive27 post says what they want to hear more of — work that challenges assumptions, celebrates student voice and explores meaningful teaching in an AI-influenced world.
A series showing the review, the pushback and the revision matches that process view and keeps pedagogy, not the logo, as the subject.
Before committing, request recent post results on comparable topics and formats, ask for audience location and role data, and agree scope, deliverables and terms directly.
For an education AI company deciding who to work with on a tool that preserves teacher and student agency, choose Mary Harper when the question is how teachers are prepared, not how a feature behaves with a learner.
Read more
Both posts focus on the systems behind the classroom: a Learning Policy Institute framework for designing teacher preparation around the science of learning, and a Texas Workforce Commission grant awarded to the Children's Learning Institute.
The grant funds CDA credentials, registered apprenticeships, associate degrees and bachelor's degrees for early childhood educators, as described from Harper's position at TXAEYC.
The preparation framework offers the kind of structure a programme proposal needs: five principles for preparation that is coherent, clinically grounded, developmentally informed, equity-centred and modelled by faculty, rather than merely described in coursework.
If the company's route into schools runs through educator-preparation programmes, accreditation conversations, clinical partnerships or a state workforce pipeline, Harper's material speaks directly to that buyer decision.
Neither post judges a product.
There is no AI feature review, account of a student or teacher making a decision inside a tool, or classroom example against which to test a prototype.
Harper's vantage point is policy, higher education and preparation systems, so a working design session would need to be built from scratch, without demonstrated product or teaching practice to draw on.
Ask for audience figures, engagement data and commercial terms before comparing reach and cost.
For the same need, Staci Lapis is the alternative when the company needs someone to build student-facing screens rather than critique the pedagogy behind them.
Read more
Their posts describe ten years of language instruction and community work before a move into UX, visual and conversational design, and a working app, Studywavve.
It lets students in different places share resources, message each other, schedule virtual meetups and exchange feedback on assignments.
The posts specify the features: an onboarding flow that auto-populates a profile card with image, language, subject, interests and location; filters for subject, interest and language; and the design systems behind the interface.
A company designing collaboration or equitable-access features, or needing a design system for an AI tool's student-facing screens, is closer to this work than to a pedagogy critique.
The 2023 post also names the areas Lapis is seeking work in, including UX, conversational design, instructional design and AI in education.
Lapis's posts describe their own app and search for a first role; they do not assess someone else's product against a teaching rationale.
There is no AI-and-agency argument to borrow: nothing about who decides when an AI support is offered, whether a learner can reject it, or how a feature changes what a teacher does next.
The posts date from October 2023 and February 2024.
Ask for audience figures, engagement data and commercial terms; a company seeking a critical read on whether a prototype protects learner judgement would gain design execution skill, but not the pedagogical test it needs.
For the same need, Dan Greenwald is the alternative when the question is how a product reaches schools, not how it treats the learner.
Read more
Both posts focus on market conditions: building a for-profit company inside systems not driven by money, where adoption is slow, budgets are tight, stakeholders are numerous and trust barriers are high; and the claim that education innovators and the ecosystems around them drive change.
In the March 2026 post on GESAwards, Greenwald says they shared insights from nearly 1,200 startups in the 2025 cycle, launched the next cycle and led a panel of seven startups scaling across countries.
It is the closest thing in Greenwald's posts to a route into the global EdTech ecosystem.
If the company wants insight into cross-country scaling, introductions to founders and educators, or a candid account of why adoption stalls, this is the layer these posts address.
Neither post addresses classroom pedagogy or student agency, the core of this need.
There is no feature review, teaching practice or account of how decisions should be shared between a platform, a teacher and a learner.
A design-review series on an AI prototype would draw on almost none of this material.
Ask for reach, engagement data and commercial terms before choosing on audience as well as subject matter.
Five LinkedIn posts by Michael Harvey cover AI and learning agency, Google's shift from Gems to Skills, the MEL26 Google Innovator Academy in Melbourne, selection as New Zealand's Microsoft Elevate Education Fellow for 2026–28, and an invited ISTELive27 session-reviewer role.
The posts do not say when they were published, so the page reflects positions across the posts rather than a dated chronology.
The LinkedIn following is about 2,900, observed on 8 October 2026; the posts do not give a role, location or engagement breakdown.
A longer reflection linked from the Innovator Academy post is not included here, and no rate card, media kit or published partnership format appears in the posts.