The objection comes up in nearly every conversation about AI-produced course materials: "Won’t it just produce the same thing as every other RTO using the same tool?"

It’s a fair concern. If you’ve tried a generic AI tool for VET content development, you’ve probably seen it happen, interchangeable workplace examples, performance criteria addressed but not embedded in any recognisable assessment structure, scenarios that feel like they were written for a hypothetical learner in no particular industry. The output is technically compliant-adjacent and completely indistinguishable from what the RTO next door would get running the same prompt.

The problem is real. But it’s not an AI problem. It’s a configuration problem.

What ASQA actually regulates

ASQA enforces the Standards for RTOs 2025 and the requirements of the Training Package. In practice, that means your materials need to address the Performance Criteria, demonstrate coverage of the Performance Evidence and Knowledge Evidence, and satisfy the Assessment Conditions in the unit of competency. That compliance layer is non-negotiable and identical for every RTO delivering the same unit.

What ASQA doesn’t regulate:

  • How you teach
  • What industry contexts you use in your scenarios
  • What your simulated workplace looks like, or what it’s called
  • How you structure assessment tasks, beyond meeting the minimum conditions
  • The tone and voice of your materials
  • The structural and pedagogical choices you make above the minimums

That silence is enormous. It’s where every RTO’s identity lives. And it’s completely outside the reach of a tool that starts from the TGA document and a blank page.

Why generic AI output is interchangeable

Most AI course development tools, and most general-purpose AI models run by internal teams, work from the same starting point: the unit of competency document and a prompt. The Performance Criteria go in, content comes out. Assessment tasks are scaffolded from the performance evidence list. The output is structurally similar to everyone else’s because the input is structurally similar to everyone else’s.

The unit document is the same for every RTO. If the configuration layer is also generic, or absent, the output will be too.

Generic AI starts with a blank page. A configured pipeline starts with your setup.

What “configured” actually means

Before Learnbuilt writes a single line of content for any engagement, we build a production environment specific to that RTO. That environment includes several distinct layers, each of which shapes the output in ways that can’t be replicated by another client running the same pipeline with different settings.

A simulated organisation built for your context. The simorg is the fictional workplace used across all project-based assessments. We build it as a privately held commercial group with multiple operating divisions – property, logistics, retail, capital works, hospitality, and corporate services. Each division operates in a distinct industry context, so when a learner encounters a project task, the scenario is set in whichever division best matches the unit’s requirements. The group doesn’t need to be in a single industry. It covers them.

This matters more than it sounds. An assessment that says “you are an employee at [company]” is interchangeable with any other assessment using the same structure. An assessment that places the learner as an employee within a specific division, using specific internal documents, responding to a specific operational scenario, that task is contextually grounded in a way that generic templates simply aren’t.

An assessment model that reflects how you actually deliver. Some RTOs are project-based, one or two capstone tasks per unit, fully simulatable, no real workplace required. Others are portfolio-heavy, drawing on learners’ own workplace evidence. Others sit between. These choices shape the structure of every workbook. We document the model during setup and apply it consistently across the full qualification.

A learner profile that informs how content is framed. Pre-employment learners need scenarios built from first principles, the simulated workplace has to do real explanatory work. Upskilling workers need the scenario to feel like a version of their existing job, not an introduction to it. These are different materials, even when they’re delivering the same content against the same TGA unit.

Your voice and terminology preferences. Some RTOs write formally. Others are direct and practical. Others are warm. Every RTO has terminology they always use and words they’ve deliberately moved away from. “Participants” not “students.” “Facilitator,” never “trainer.” These are small details, but they’re what make materials feel like they were written for your RTO rather than retrofitted to it.

The simorg as the single biggest differentiator

Of all the configuration inputs, the simulated organisation has the most visible effect on how materials read. It’s also the element that most clearly illustrates why custom configuration can’t be faked with a generic prompt.

The diversified commercial group model exists for a specific reason: a business with multiple operating divisions in genuinely different industries can provide authentic scenario variety without rebuilding the organisation for each unit. When a unit’s scenario requires a supply chain context, it’s set in the logistics division. When the next unit requires a retail context, it moves to the trade supply division. The group’s core structure, its leadership, its shared policies, its finance and HR functions, stays consistent across the entire qualification. Learners build familiarity with the organisation over time rather than starting from scratch each unit.

That consistency also makes the simorg documentable. Policies, org charts, email chains, meeting minutes, project briefs, all of it lives in a persistent fictional organisation that grows as production progresses. Any document referenced in an assessment task has to exist in the simorg. That’s a hard requirement, and it’s one that generic tooling can’t satisfy.

What this looks like in practice

Two RTOs could both use Learnbuilt to build BSB50420 Diploma of Leadership and Management. One delivers primarily to healthcare team leaders in metropolitan hospitals. The other delivers to small business owners across regional Australia.

The TGA compliance layer is identical. The assessment conditions, performance evidence, knowledge evidence, all the same. Neither RTO gets to change any of that.

Everything else is different. The simorg’s client in each scenario. The industries drawn on for examples. The assumptions built into the scenario about the learner’s context. The tone. The structural choices that reflect how each RTO actually assesses and how their learners actually engage with materials.

That differentiation isn’t produced by asking the AI to “write for an RTO that delivers to healthcare workers.” It’s produced by a production environment that was configured with that context from the start, and that applies it mechanically through every stage of the pipeline.

The question worth asking

The right question about any AI-produced VET materials isn’t “was this generated by AI?” Auditors don’t ask that, and it’s not what ASQA assesses.

The right question is: does this reflect how our RTO actually operates, how our learners actually engage, and how we actually assess?

If the answer is no, the problem isn’t the technology. It’s the inputs. The compliance layer will always look like everyone else’s, it has to, because it’s derived from the same Training Package document. Everything surrounding that layer is configurable, and with the right setup, the output is as specific to your RTO as anything a skilled human writer would produce.

It just takes considerably less time.