Australian corporate training teams that started mixed reality pilots in 2023 and 2024 are now in their second or third generation of programme design in 2026. The shape of what is working is clear enough to write about as a practitioner pattern rather than a vendor pitch.

The programmes that have stuck:

Safety induction for industrial and resources environments. Mining, oil and gas, heavy construction, and rail operators are running MR-based safety inductions for new hires and contractors. The MR programme runs through specific scenarios — confined space entry, working at heights, plant lock-out tag-out, emergency egress — with the trainee performing the actions in the headset. The completion data is measurable. The reduction in low-grade incidents in the first six months on the job is measurable. The training compliance argument closes.

Clinical procedure rehearsal for Australian health services. State health services have built MR libraries for nursing procedures and selected medical procedures. The trainee rehearses in MR before the supervised clinical encounter. The supervising clinicians report the trainees are more prepared, the patient encounters are faster, and the learning curve is shorter. The investment is paying back on a workforce capacity metric, not just on a training cost metric.

Equipment maintenance training for utilities and infrastructure operators. The maintenance technician learns the procedure on a virtual representation of the asset before being supervised on the physical asset. The cost of asset downtime for training has come down. The cost of training itself has not changed much but the throughput has improved.

What has not stuck:

Sales training and soft-skills training in MR. The headset adds friction that the workflow does not need. Soft-skills training has moved back to video-conference-based formats with AI roleplay assistance, which is a different technology stack from MR.

Leadership development in MR. The early pilots did not produce sticky behaviour change and the budgets moved elsewhere.

General-purpose product training in MR. The “use MR for everything” rollout strategy did not survive the second year of cost review.

The budget shape that is working in 2026:

A capital line for the headset fleet, with a three- to four-year replacement cycle. The newer Quest generation hardware is cheap enough that the fleet conversation is not the budget conversation it was in 2022.

A recurring line for content production, either with an internal team or a contracted content partner. The content cost is the cost. The headsets do not produce the content; people do.

A platform line for the MR learning management system and the analytics. The training teams running serious MR programmes have moved off the bundled vendor LMS and onto specialist MR learning platforms with better analytics.

For Australian training teams designing the next round of MR investment in 2026, the read is that the technology is mature, the content production discipline is the rate-limiter, and the programmes that pay back are the ones with a measurable workforce outcome at the other end. For training teams that want to integrate AI-assisted personalised learning into the MR programme — adapting the training scenario to the trainee’s prior performance — the work moves into the AI implementation conversation. Team400 is one of the Australian AI consultancies working on the AI side of corporate learning programmes.

The 2026 read is that MR is past pilot, the use cases are clear, and the training teams who took it seriously through 2024 are now running a craft. The training teams who waited for the technology to mature have a shorter on-ramp ahead of them.