Every RTO that looks at Apollo asks the same question sooner or later, usually in the compliance meeting: who made the judgement? It’s the right question. An auditor is entitled to ask it, and the answer decides whether a marking tool belongs in a training organisation at all.
So the name isn’t branding. It’s the product description, three words long, and each of them is doing work.
Assessment. It reads the real evidence
Apollo works on what the student actually submitted, in the form they submitted it: Word documents, PDFs, project templates, portfolios, even scanned pages. Each answer is read against the task brief that set the situation, the marking guide that states the requirements, and the sample answers that show the expected standard. If it’s not on the page it doesn’t get credited, and if it is on the page it doesn’t get missed.
Judgement. It belongs to a person
Under the Standards, a qualified assessor decides whether a learner is competent. Apollo is built so that it can’t. It has no authority to finalise an outcome and no way to get there without a person clicking. It drafts the marking, lays the ticks and crosses where the evidence sits, attaches its reasoning, and stops.
Nothing is final until a person makes it final.
Assistant. It does the hours in front of the decision
The reading, the cross-referencing against the marking guide, the drafting of feedback, the placing of marks. At one college that work came to 16.25 minutes an assessment by hand and 5 with Apollo. That’s the time in front of the judgement, not the judgement itself. The assessor arrives at a page that is already laid out and spends their attention where it belongs: on the borderline calls.
The returned document: the assessor’s name, the time of sign-off.
Why not “AI marking”
“AI marking” would be easier to sell, and it would describe a different product: one that puts your RTO on the wrong side of the question this article opened with. Plenty of platforms will mark for you, as long as the work is typed into their question fields. The name isn’t a disclaimer bolted on afterwards. It’s the design constraint the software was built around, and the reason there’s no button anywhere in Apollo that finalises an outcome on its own.
And the record says so
Every returned document carries the assessor’s name and the time they signed it. If a validator asks who made the call on a borderline result, there’s one answer and it’s a person. That is a narrower promise than marking, and a much easier one to defend.