How do you know classroom tech is actually helping?

A researcher, a principal and Atomi’s co-founder on how to tell whether classroom technology is earning its place.

Next year, the screen time debate moves into the classroom with NSW public schools being asked to limit screen use in class from Term 1 2027, becoming mandatory soon after in 2028. Planned device-free time in secondary schools across Victoria will also come in 2027, with policy setting rules for schools, expecting teachers and faculty to figure out what technology stays and what goes.

Most faculties didn’t choose their platforms all at once, rather they collected them: a trial here, a mandate there, a demo that promised to save time. Often the question of whether the technology actually helped students in the classrooms was unanswered.

In September, we put that question to Queensland principals and deputy principals at Atomi’s Beyond the Bell, an evening we hosted with the Queensland Secondary Principals’ Association (QSPA). Dr Tony McCormack, Executive Principal of Woodcrest State College, sat alongside Queensland University of Technology researcher Dr Julie Arnold and Atomi co-founder Rob Barakat. Sarah-Eleni Zaferis, a former teacher and now Product Enablement Specialist at Atomi, moderated.

The question we asked was: “How do you know if the technology is actually helping?”

Start with the learning, then look for a tool

“You’ve got to be driven by the purpose. What is the learning that can happen as a result of this tool that couldn’t happen in another way?”
Dr Tony McCormack, Executive Principal, Woodcrest State College

If nothing digital passes that test, he’s comfortable with that. “If a piece of butcher’s paper and a pen is the best tool for the job, then that’s what you use.”

Julie’s research explains why the test works. What moves outcomes, she said, is how teachers and students interact around evidence of learning: the purpose of the lesson, what quality looks like and how students respond to feedback. “Those interactions around evidence are the thing.”

Rob put much of the current frustration down to volume. “Can we do more with less? Can we do more with fewer platforms?” It’s a quick audit for any faculty. List every platform and the learning each one makes possible. If two do the same job, or one does nothing, it goes.

Define success before the tool goes in

You can’t know a tool is helping if nobody agreed what helping means. Rob said the schools that get the most from technology have “a pretty clear idea of what they’re trying to achieve”. Broad goals like lifting results or general engagement are “a road to failure,” he said, “because it’s very difficult for us to plug that gap.”

Tony framed it as “that vexed question about measuring what we value versus valuing what we measure”. A learning intention on the board is easy to check, but in a compliance culture it says little about “the actual true learning that we’re looking for”.

So before a tool goes in, agree with your faculty what students should be able to do differently if it works.

Look beside the screen for evidence

Logins aren’t learning. Usage data shows a student opened the platform, not whether they were thinking through the content. Julie pointed leaders away from the tool and towards the classroom around it.

“Evidence of students explaining, having ideas and interacting […] The signals are not best in the tool, but they’re adjacent to the use of the tool.”
Dr Julie Arnold, Researcher, Queensland University of Technology

The signal is one any head of department can check on a walkthrough. “Can that teacher return to that learning goal? Are students actually talking about what success looks like in relation to that goal?” If students can explain what good work looks like, the tool is probably doing its job. If they can only tell you what they clicked, it probably isn’t.

Tony had a warning for anyone hoping to turn that into a rubric. “To try and narrow those complex problems down to a simple solution that’s easily replicable, that can be ticked off on a check sheet, is a fool’s errand.”

The research supports a cautious approach. Stanford’s 2026 review of AI in schools found adoption is running ahead of the evidence. The EEF’s 2025 review found edtech’s effects are mixed, and weaker for disadvantaged students. A 2025 consensus study in Humanities and Social Sciences Communications argues edtech should be judged on evidence of learning, not on adoption. Until that evidence catches up, the most reliable data you have is what students say and do in your own classrooms.

Next year’s decisions

Term 4 is when next year’s technology decisions get made, often in a rush. So we closed the night by asking each panellist for one thing a school leader could do first:

• Julie would start with teachers and one question: “Is this tool […] producing useful and interesting evidence of student learning for me, or is it making my already complex job even more complicated?”
• Tony would talk to the people the change is for, then help them find the solution themselves. “We got to walk alongside them and make sure that that ownership is there and plant the seed with somebody because they might not yet understand or know what the solution to their problem is.”
• Rob told leaders to ignore the shiny features and “work back from intent”. Then hand the hard part to the vendor: “You’ve got to put the onus on them to prove to you how their solution actually works in lock step with that.”

We’ll take that test too. Tell us what your faculty’s tech needs to achieve. We’ll show you whether Atomi gets you there. Book a demo →

“As the technology gets more powerful, the role of the teacher becomes more important, not less important.”
Rob Barakat, Co-founder, Atomi

Sarah-Eleni closed the night with the line every answer had come back to: “The tool is never the point. What you’re trying to achieve for students is what actually matters.”

Before your faculty locks in 2027, take one question to your next meeting: for every tool we use, can we name the learning it makes possible that wouldn’t happen any other way?