From ad hoc to aligned: TAFE Gippsland’s approach to AI in learning content design

TAFE Gippsland wanted educators to use AI safely and consistently to build better learning content faster. Through the FSO Skills Accelerator-AI, it partnered with RockMouse to pilot SPARK, an AI assistant that turns training requirements into structured learning materials that educators review and approve.

At a glance

FSO Skills Accelerator-AI initiative: AI Learning Content Assistant pilot

Partners: TAFE Gippsland, RockMouse

Timeframe: Skills Accelerator Action Learning Sprint, 8 weeks in late 2025

Audience: Educators at TAFE Gippsland

Primary goal: Develop an AI assistant that generates structured, reliable learning content aligned to Units of Competency (UoCs)

Key outcome

By embedding AI into educator workflows, TAFE Gippsland has lifted capability, streamlined processes, and created a scalable model for developing high-quality learning materials.

The challenge

TAFE Gippsland educators faced significant time and workload pressure when developing learning materials, including learner resources, activities and lesson plans aligned to UoCs.

Educators were already using AI, but in different ways and with different levels of confidence. This created three risks: inconsistent quality, limited visibility of sources and prompts, and possible misalignment with training requirements.

AI‑generated content also created practical issues when copied into Moodle, including hidden formatting and metadata that added rework for the eLearning team.

TAFE Gippsland needed a consistent, governed way to use AI that still kept educators in control.

About the initiative

The pilot developed SPARK, a learning design assistant that guides educators through a structured, step‑by‑step process for designing learning materials. Educators upload unit documents, training and assessment information and delivery context, then use the assistant to generate draft learner resources, activities and session plans.

Educators remained in control throughout. They reviewed, contextualised and validated all AI‑generated content before it was used with learners.

The pilot was grounded in staff research. In a survey of 36 staff, including 31 educators, just over half (51%) said they used AI for curriculum‑related tasks at least weekly, mainly to create learner activities, generate resources and write lesson plans.

The impact

Testing of SPARK with nine VET educators showed that AI can support learning content development when it is integrated into educator workflows and governed by human review.

Faster content development

The assistant streamlined the development of learner resources, activities and session plans aligned to units.

More consistent outputs

The tool helped create clearer structure, stronger source control and cleaner outputs for Moodle and other platforms.

Stronger educator confidence

The pilot supported responsible AI use while keeping educators accountable for review, judgement and context.

What worked well

Built around real workflows

The assistant mirrored how educators already develop learning content, making it more relevant and easier to test.

Flexible delivery

The project team adapted as user needs and AI capability evolved during the sprint.

Better outputs as models improved

Newer AI models enhanced the structure and reliability of draft learning content over the course of the sprint.

The strongest lesson was simple: AI tools work best when they fit the way educators already work.

“This initiative demonstrates how a governed AI approach can significantly reduce the effort required to develop learning content, while improving quality and consistency across delivery.”

Craig Simon, CEO & Co-Founder, RockMouse

“We need to lean on AI a bit more and let it do the bulk of the lifting. It’s the same as someone moving a pallet of bricks by hand when there is an electric pallet jack right next to them.”

Anonymous survey comment, TAFE Gippsland

Challenges faced

Shared language mattered

VET terms, assessment language and internal processes had to be clearly defined so the tool could reliably produce useful outputs.

The technology kept moving

Improvements in AI capability meant early prototypes had to be reworked, which extended development time but improved the result.

Key insight

The pilot showed that AI projects in VET need a shared language, close testing with users and room to adapt as the technology changes.

Next steps

TAFE Gippsland will now test the AI assistant in live delivery. Feedback from educators will be used to improve usability, output quality and alignment with institutional standards. Longer term, the project will explore deeper integration with Moodle and Training.gov.au, and the potential to scale the model across the organisation.

Share this article