On 5 November I sat on a panel at EduTech Asia 2026 in Singapore. The conference brought around 8,000 educators and 250+ speakers to the Sands Expo & Convention Centre at Marina Bay Sands over two days. Our session was in the K-12 Teaching & Learning stream, titled "Beyond the scoreboard - rethinking how we measure student success". Jeremy Birk from Wellington College China moderated. On stage with me were Barbara Tajti, Secondary Principal at Saint John Mary International School in Thailand, and Adrian Lim, Head of Innovation at M.Y World Preschool here in Singapore.
We had 40 minutes, so we went straight for the tensions. Here is roughly what we said.
We say one thing and count another
That was the line that framed the session. Schools write mission statements about curiosity, resilience and collaboration. The only numbers on a student's record, though, are grades and exam scores.
The gap is measurable, not just rhetorical. 83% of educators say they emphasise durable skills, but only 24% can actually measure them (ETS, 2026). Underneath that sit the engagement numbers: 81% of highly engaged students feel prepared for the future, against 27% of the least engaged, and a third of students don't even feel they belong at school (Gallup, 2026).
Nobody on the panel argued that exams should go. The argument was about what sits next to them.
What counts as success
My segment: measuring what matters versus what is easy to measure. Evidence of durable skills is not a single test. It is a portfolio (artefacts, structured tasks, observation, student reflection, collected over time). Melbourne Metrics' "ATAR+" learner profiles that sit alongside the ATAR are a working example of that. Mastery transcripts are further along (370 member schools, 500 colleges accepting them). Real, but still niche.
Exams and admissions
Barbara took this one, and she is right that external exams have a gravitational pull nothing else in education has. England is trimming GCSE exam time by roughly 10% while keeping exams as the backbone. The IB's 16+ review adds choice and wellbeing while tightening exam security at the same time. Dartmouth, Yale, Brown and Harvard have all reinstated the SAT/ACT. The realistic model, in her words: protect exam outcomes and add a richer evidence layer. Supplement, don't replace.
Parents and the story
Adrian's point was about framing. Broader measures are not a retreat from rigour: present them to parents as better evidence of readiness for top universities, and the conversation changes. Singapore's own removal of mid-year exams showed the flip side too: take away a familiar signal without giving families a readable replacement, and you get anxiety, not relief.
AI, data and evidence
Then back to me. AI has broken product-only assessment. If you only assess the product, the product is what gets generated. AI-use policies (the discursive fix) create an illusion of enforcement. Structural fixes produce evidence you can actually trust: process documentation, oral defences, tasks redesigned so they cannot be delegated to a chatbot.
The upside of AI is continuous formative assessment instead of a few high-stakes moments. The risk is drifting into surveillance-style data collection. PISA 2025 gives a decent guardrail: moderate device use was linked to stronger outcomes, excessive use to weaker ones.
The scoreboard isn't going away, and it shouldn't. But a student's record should tell the rest of the story too. Right now, in most schools, it doesn't.