What Teachers Actually Need from AI
The promise of AI in classrooms keeps getting framed around student outcomes. But the bigger leverage point might be the fifteen minutes of prep teachers do for every hour of teaching.
Writing from the partners on education, AI, and the infrastructure of knowing. We publish here because thinking in public is part of how we do diligence — and because founders deserve to understand how we think before they take a meeting.
The promise of AI in classrooms keeps getting framed around student outcomes. But the bigger leverage point might be the fifteen minutes of prep teachers do for every hour of teaching.
We're in the early stages of rebuilding how institutions store, transmit, and certify knowledge. The companies that get the plumbing right will be the dominant infrastructure platform for learning.
Education Perfect. Kami. Learnosity. Three NZ-originated companies that found global product-market fit. What do their trajectories tell us about the structural advantages of building EdTech from the Pacific?
Assessment isn't just evaluation. It's the signal layer for every adaptive system in education. Whoever owns the assessment API owns the feedback loop that every AI in education depends on.
Project-based and work-integrated learning consistently outperform lecture formats on long-term retention. Until recently, the bottleneck was facilitation bandwidth. AI changes that equation.
Collecting engagement signals on children is one of the most consequential data questions in consumer tech. Founders who treat it as a compliance box are building on sand.
New Zealand and Australia have high-trust education systems, strong R&D culture, and institutional buyers who are unusually willing to pilot new tools. That combination is rare globally.
Every knowledge worker leaves a trail of work across platforms that disappears. The tools that help them own and organise that trail have a $30B addressable market nobody is talking about.
After two years and ten investments, we've developed a working thesis on what separates EdTech companies that scale from those that stall at pilot-stage.
Before Rimu, I spent five years at the firm that built the most-deployed LMS in Australasian schools. The things that surprised me about institutional education buyers have shaped every investment thesis I hold now.
Degrees are expensive partly because they're scarce. The scarcity is regulatory, not pedagogical. The companies that figure out accreditation infrastructure will unlock a market worth hundreds of billions.
Curriculum development is a $12B professional services market in the English-speaking world. It runs on Word documents and institutional knowledge. That's the setup for a software wedge.
Reading acquisition is one of the most consequential and measurable outcomes in primary education. It's also one of the few places where adaptive assessment software has proven, peer-reviewed evidence of efficacy.
Duolingo captured the consumer narrative, but conversational AI tutoring is opening an entirely different market: professional language development for enterprises with multilingual workforces.
Every EdTech vendor claims 'personalised learning.' Most means showing different content based on a pre-test. Real adaptive learning is a feedback loop that changes the instructional strategy, not just the difficulty.
Eighteen months into deploying Rimu Fund I, we wanted to write down what we've learned about where AI is creating durable value in education — and what we still don't know.
Most EdTech products treat assessment as a feature. The companies that treat it as a platform — an API that other products build on — have a fundamentally different competitive position.
Software ate the workflow. AI is eating the knowledge layer. The difference is consequential: knowledge tools need to be trusted, curated, and owned — not just used.
Capital for seed-stage EdTech in NZ and Australia has historically been thin relative to the quality of founders. That's starting to change — and the composition of what's getting funded is shifting toward AI-native tools.
Education has a longer institutional sales cycle, a more complex buyer landscape, and a higher trust requirement than almost any other market. Founders who have lived inside that system build differently.
New Zealand's education system has historically been ahead of the curve on formative assessment practice. The arrival of AI-powered assessment tools is accelerating changes that have been building for a decade.
Formative assessment data — the small, frequent checks teachers run throughout a lesson — is the richest signal available to adaptive learning systems. Most EdTech products ignore it entirely.
Rimu Capital closed its first fund in June 2023. Here's what the first year of operating a seed-stage EdTech fund from Auckland has looked like — and what we'd do differently.