Where AI Is Headed in the Next Five Years, and What That Means for RTOs

AI will not replace the RTO. It will sit inside the same work you already run: enquiries, enrolment, training, assessment, student support, and reporting.

The useful question is not “what model will win.” It is “which parts of the college will be drafted by a system, which parts still need a qualified person, and who is accountable when the output is wrong.”

Nobody can give you a certain five-year map. Tools will change. What is already visible is the shape of the work. This is an outlook, not a forecast you should treat as fact, and it is not legal advice.

What Is Already Locked In

Three things will not go backwards.

Students will keep using AI. They already draft, summarise, and practise with it. Pretending they do not is not a strategy. The job is assessment design and integrity, not a blanket ban you cannot enforce.

Admin will keep absorbing AI first. Acknowledgements, reminders, routing, first-pass summaries, and weekly visibility are where the time sits. That is still the safest place to start.

Regulators will keep asking about quality, not novelty. ASQA’s principles for responsible AI use in VET do not add a new Act. They tell you to apply the Standards you already have: governance, human oversight, privacy and records, equity, and alignment with the training product. That direction is more likely to tighten in practice than to disappear.

Where the Technology Is Actually Going

Over the next few years, expect less “one chatbot on the website” and more small systems that do one job in a workflow.

You will see more agents that draft and hand off, not oracles that decide. You will see more connection into the tools you already use (CRM, inbox, student management system) rather than a second college living in a chat window. You will see more logging: what was asked, what was drafted, who approved it.

You will also see more noise. Vendors will sell “AI campuses.” Most of that will be the same enrolment leak with a new interface. The providers who do well will be boring on purpose: one path, approved facts, a named human, a record you can show.

Models will get cheaper and more capable. That does not make them safe to point at competence decisions, visa questions, or fee balances. Capability and permission are different things.

What This Means for Assessment

Assessment is where the next five years get hard.

If AI can produce a polished written answer, written-only evidence gets weaker. Expect more weight on observation, workplace evidence, professional discussion, and tasks that are awkward to fake. Simulation and AI practice tools will spread. Using them as the only evidence of competence will stay a validity problem.

Trainers and assessors will spend less time formatting materials and more time judging whether the evidence is authentic, sufficient, and mapped to the unit. That is not less work. It is different work. Staff capability becomes part of quality, not an optional extra.

What This Means for Enrolment and Operations

The enrolment office will look more like a controlled pipeline and less like a shared inbox.

After-hours coverage will become normal, not a nice-to-have. Document chasing will be timed and visible. Leadership will expect a weekly pulse without a Friday spreadsheet. The student management system will still be the official record. New tools will sit beside it, read first, write only where you have agreed rules.

Education businesses that are not RTOs (short courses, coaching, private institutes) will feel the same pressure on speed-to-lead and follow-up. They will not have the same Standards, but they will have the same reputational risk if a system invents a start date or a price.

What This Means for Students and Fairness

If a tool only works for digitally confident students, or mishears accents, or has no human alternative, it will create a two-speed college.

Over five years that gap will show up in completions and complaints, not in a launch blog post. Build the alternative path now. Test with the cohort who will struggle first. Principle 4 in ASQA’s set (equity, inclusion, accessibility, wellbeing) is not a slogan. It is whether people can still demonstrate what they can do.

What Not to Bet the Next Five Years On

Do not bet on one vendor, one model, or “AI will mark the assessments.”

Do not bet on replacing your student management system. Do not bet on a policy PDF that nobody runs. Do not bet on hiding student files inside consumer tools because it is convenient.

Do bet on three capabilities you can take from tool to tool: a map of your real workflows, a knowledge base you own, and a habit of human approval on anything that commits the college.

A Sensible Posture from Here

Use the next year to get one workflow right: usually enquiry response or incomplete files. Use the years after that to widen, with the same rules.

Keep a register of tools. Keep decide jobs with qualified people. Keep the official record in the system of record. Read ASQA’s live guidance when it updates.

Five years from now the brands on the invoice will be different. The test will look the same: did the student get a fair, defensible outcome, and can you show who was responsible?