-11.jpg)
How are RTOs really using AI? At aXcelerate Day 2026, three VET leaders gave a practical answer. Most RTOs already use AI for everyday writing tasks, fewer use it for content development and analytics, and the providers getting the most value start with one small problem, keep people in the loop, and put governance in place before adopting new tools.
Amina Jackson, Manager, Customer Success at aXcelerate, opened the session with a shift most people in the room recognised. In the past 1-2 years, AI in RTOs has gone from "wow, this is amazing" to "we need a strategy for this." The questions had changed too. RTOs were no longer asking whether to use AI, but where it saves time, what the risks are and how to put the right guardrails around it.
The three speakers who followed answered those questions from different perspectives. Chris gave the sector-wide view, Mark showed what a well-chosen project can deliver, and Shane showed what it takes to build a framework inside a business.
Chris Enright of RTO Logic and RTO Radar has more than 18 years' experience in the VET sector, working with RTOs on strategy, compliance, operational improvement and capability. Much of his recent work has focused on what responsible AI adoption looks like for providers. He and his team speak with 50-100 organisations and people in the sector every week, which gives him a wide view of how AI is really being used. By his observation, every one of them is using it in some form. Around 80% use it for everyday writing such as emails and policy documents, around 30% for content development, and around 15% for analytics such as monitoring student progress and triggering interventions. A few years ago, most of those numbers would have been close to zero.
He sees two clear reasons RTOs are adopting AI. The first is efficiency: reducing operational costs and improving accuracy. The second, and the one he's most enthusiastic about, is better service delivery. The time AI saves can be put back into areas that create more value for students.
The conversations have shifted too. A few years ago, RTOs asked him what AI was and whether they should allow it at all. Now they ask how to use it to work more efficiently, and how to protect themselves from the risks. That second question matters because, in his experience, at least three in four organisations have staff using personal AI accounts, often without any formal review of the tools involved.
For Chris, the answer is a deliberate approval process. He walked through ASQA's Principles for the Responsible Use of AI in VET, which cover governance, human oversight and accountability, privacy and data security, equity and accessibility, and alignment with training package and industry needs. Then he turned it into a challenge. RTOs already have documented processes for training and assessment strategies and continuous improvement registers. They should have one for deciding which AI tools are allowed in, too.
He suggested RTOs work through a few key questions before approving any AI tool:
"These are the questions that RTOs now need to start asking themselves," Chris said. "Not, should I bring AI into the building."
In practice, AI governance for an RTO comes down to three things: a documented process for approving tools, clarity on who uses them and how, and monitoring once they're in place.
Mark Costello, Managing Director of Asset College, has worked in the VET sector for more than 20 years and founded the business that became Asset College in 2006. Asset delivers more than 50 qualifications across Australia and recently won Large Training Provider of the Year at the 2026 Queensland Training Awards, making it a national finalist at this year's Australian Training Awards. His team is working on dozens of AI applications across the business. Rather than walk through them all, Mark told the story of one, because that's where he believes every RTO should begin: find a small problem and make it easier to solve.
For Asset College, that problem arrived in the mail. Much of its regional and remote delivery runs on paper, and completed enrolment forms and records from weekend courses land on a staff member's desk each week. Uploading them meant scanning, renaming, searching for each student and attaching every file by hand. It took them five to six hours.
Their team built an app that connects to aXcelerate through its API. Staff pull a class list, print a QR-coded cover sheet for each student and scan the whole pile at once. Each file goes to the right student portfolio, and AI reads the handwritten enrolment forms and maps the details to the student record. Anything the app can't match goes into a review queue, where staff check the extracted data against the scan before submitting. It's the human oversight Chris described, built into the workflow. Security was considered from the first planning session, and the team now runs its AI in-house to keep data on its own systems.
After a few early bugs were ironed out, the weekly work took about 45 minutes.
"Start with something small," Mark said. "Do something that will save you a few minutes a day."
For RTOs new to AI, a single, well-defined workflow with a human review step is a practical and low-risk place to begin.
Shane Darwin is General Manager of JB Hunter, a specialist telecommunications training provider that has been around for 35 years and has a long history of trying new technology early, from online training to virtual reality. So when AI arrived, the business was keen to explore it. Around 18 months ago, the team began building an AI agent to answer incoming calls. When JB Hunter's compliance manager came back from a seminar urging the business to take AI seriously, Shane took it as a sign it was time to go further.
The team arranged meetings with four AI providers in a single week to learn what AI could do for an RTO. The providers offered to build custom agents, typically for a few thousand dollars each, including one that could help JB Hunter manage its work-based learning data in aXcelerate. One provider couldn't make a meeting and replied by email instead. He offered to build the agent for a fraction of the others' prices, at no upfront cost, and asked JB Hunter to send its API credentials to get started.
It seemed too good to be true, and Shane recognised it for what it was. Handing over API credentials would give an unknown third party direct access to student records, so the offer went no further. For Shane, it confirmed that JB Hunter needed a clear process for assessing AI providers before any conversation reached that point.
Around the same time, Shane needed to draft an email reply to a student and turned to his personal AI account to help write it. Then he stopped to think. "If this is me doing that, what are my staff doing?"
Those two moments became the start of JB Hunter's AI framework. The business set up an AI committee made up of the general manager, compliance manager, systems manager and a member of the sales team, which meets weekly. The team drew on the federal government's AI resources at business.gov.au and ASQA's guidance to write an AI policy. They also created a tool register that every new tool must pass through before it's approved. It asks the same questions each time: where the data is stored, what is kept by the AI model, and how the tool will be used in training and assessment.
"One of the very first questions we ask now is: where's the data, where's it being stored, and how secure is it?"
Shane also rethought how AI tools should be chosen. Selecting a Student Management System is a long-term commitment. AI tools change too quickly for that, so JB Hunter now prefers short commitments, monthly where possible, so it can try something and move on if it doesn't work.
That flexibility has already been tested. The AI phone agent the team started with knew the business well but asked callers too many questions, which lengthened the time before they reached a person. Rather than switching it off, JB Hunter moved it to after-hours and weekend calls, where its knowledge is useful and speed matters less. The committee also found the business was paying for AI subscriptions nobody was using, and now reviews usage and value as a standing item.
The framework has produced wins too. The team built an AI agent that draws data from aXcelerate's Work-Based Learning feature through the API and turns it into clear, reportable information. The admin team now uses it to track and sign off each student's progress.
Heard together, the three talks told one story: the principles, a project that proves them, and a framework that holds them in place. Each speaker came back to the same few ideas: decide deliberately before you adopt, know where your data lives, keep people in the loop, and review what's delivering value and what isn't.
The principles the three speakers shared are the same ones guiding AI at aXcelerate: maximising value and minimising vanity. The focus is on AI features that solve real problems for RTOs, built with close attention to data security and human oversight.
Several of those features are already in aXcelerate. AI assessment question generation turns an uploaded document into draft assessment items. AI-assisted marking supports trainers and assessors with short answer and multi-answer questions. Behind the scenes, aXcelerate's AI features run on Australian-based infrastructure and respect existing user permissions, so data stays onshore and people only see what they're authorised to see. That foundation also supports the new AI Assistant, now available in closed beta. It helps you find and answer questions about the data in your account, and also gives help with aXcelerate features and functionality. Its capabilities are growing, and you can ask it "What can you do?" for an up-to-date list.
Want to learn more about AI in aXcelerate? See our AI features.
Q: How are RTOs using AI?
A: Many RTOs use AI for everyday writing tasks such as emails and policy documents. Some RTOs use it for content development or for analysing data to monitor student progress. RTOs using aXcelerate can also use AI built into the platform, such as assessment question generation and AI-assisted marking, while others are building their own automations, such as processing paper enrolment forms.
Q: What does ASQA say about using AI in VET?
A: ASQA has published five Principles for the Responsible Use of AI in VET, covering governance, human oversight and accountability, and the safe, ethical use of AI. They don't add new regulatory requirements, but help providers apply existing obligations under the 2025 Standards for RTOs. Decisions affecting students must remain the responsibility of qualified trainers, assessors and staff.
Q: How should an RTO start with AI governance?
A: Start with a documented process for reviewing new AI tools, including where data is stored, who will use each tool and how its use will be monitored. A dedicated committee and a tool register can give that process structure.
Q: Is it safe for RTOs to use AI with student data?
A: It can be, if the RTO knows where data is stored, who can access it and what the AI tool retains. Tools built into a Student Management System can be simpler to govern, because they sit within existing security and user permissions. aXcelerate's AI features run on Australian-based infrastructure and respect user permissions.
Q: What AI features does aXcelerate have?
A: aXcelerate includes AI assessment question generation, which turns an uploaded document into draft assessment items, AI-assisted marking for short answer and multi-answer questions, AI template generation for on-brand communications and more. The AI Assistant, now in closed beta, helps users find and answer questions about data in their account and get help with aXcelerate features. Talk to the aXcelerate team to see how they could work in your RTO.
Q: Where is aXcelerate's AI data hosted?
A: aXcelerate's AI features run on Australian-based infrastructure, so data stays onshore. Read aXcelerate's AI Principles to learn more.
As 2025 draws to a close, we’re taking a moment to reflect on an extraordinary year of transformation, innovation and partnership within the aXcelerate community.
A solid way of achieving high learner engagement online is through your LMS, which should be able to provide the ability to author online content, lessons and assessment with interactive question types.
