Role plays and branched scenarios are among the best ways adults learn. But they take time and resources many providers do not have. Can AI help more students get the practice they need in high-stakes situations before they enter the workplace?
At a glance
FSO Skills Accelerator-AI initiative: AI role-play pilot
Partners: Canberra Institute of Technology, Phoenix Academy, GenLearn
Timeframe: Skills Accelerator Action Learning Sprint #3, March–May 2026.
Audience: Early-stage students practising challenging workplace conversations.
Primary goal: To test if AI can provide students with role-play practice scenarios to improve realism, confidence and industry readiness in high-stakes situations.
Key outcomes
- Increased student confidence across key communication skill areas
- Improved readiness for live role-play assessments
- Less trainer and assessor time needed to provide the same level of support and practice
- A projected 15% lift in course completion rates at Phoenix Academy.
The challenge
Time, budget and limited opportunities to practise newly developed skills continue to affect student outcomes.
This matters most in fields such as community services, healthcare, HR, management and frontline support, where workers regularly need to handle challenging conversations with judgement, empathy and clear communication.
Many students start their training with limited confidence in applying these skills in real time. VET delivery does not always provide enough structured practice time before assessment or employment.
Classroom role-plays can help, but they can feel high pressure, especially for early-stage students with English as an additional language.
This pilot explored whether AI-supported role-play could give students a safe, repeatable space to practise realistic workplace scenarios aligned to unit performance criteria and knowledge evidence.
About the initiative
Through a Skills Accelerator Action Learning Sprint, the initiative brought together Canberra Institute of Technology, Phoenix Academy and GenLearn to pilot a role-play tool.
The tool helps educators rapidly turn their own course materials into interactive AI instructors, giving students the repeatable, real-world practice they need to build competence and confidence.
GenLearn provided the technology solution and mentoring support. The training providers tested how the tool could be applied in their own classrooms.
Phoenix Academy
The academic team built AI chatbots to help students recognise tone, assess different situations and practise appropriate responses. Students used the tool to practise scenarios including:
- Engaging with individuals who may be distressed or averse to change
- Communicating complex information to diverse audiences
- Adapting communication style to deliver sensitive information
The scenarios aligned to ‘Lead communication in the workplace’ (BSBXCM501). They were delivered face-to-face with a diverse cohort of 10 international students studying for a Diploma of Business (BSB50120).
Canberra Institute of Technology
CIT ran the pilot in its Community Services programs, with a focus on formative learning for early-stage students. More than 30 students participated in the pilot. The tool was designed to complement, not replace, face-to-face role-play. It provided:
- AI-generated client scenarios reflecting workplace situations
- Real-time conversational interaction using voice
- Automated feedback on communication approach and structure
- Flexible, self-paced practice with repeat opportunities.
The pilot also considered what would be needed to scale the approach, including integration with their learning management system, data privacy, compliance and cost management.
The impact
The pilot showed that AI-supported role-play can help students build confidence and workplace readiness, while also making delivery more efficient for trainers and educators.
Phoenix Academy reported:
- a 15% lift in expected course completion
- 100% positive feedback from staff and students
- a 50% reduction in the time needed to provide the same volume of practise activities
- a 70–80% reduction in the time needed to deliver contextualised feedback
CIT showed that AI role play can fit within existing teaching models and support students with different starting points, including beginners from diverse language backgrounds.
Across both trials, students consistently described the tool as a safe space to practise, make mistakes and improve without the pressure of assessment or being observed by peers.
“The one outstanding comment I heard today was if they had had this technology in their early stage, they would have walked into role plays far more confidently compared to the original methods that we’ve been using so far.”
Meg Mills, Senior Educator, CIT Community Work
“The sprint project provided tangible evidence that the human component remains integral for ethical AI use, utilising its original purpose as a supportive ecosystem for research, comparison, and extended practice. By positioning AI as a resource for the trainer rather than a replacement, we don’t just protect the integrity of learning, we ensure AI is used ethically and intentionally; not as a substitute for human guidance, but as a powerful tool to deepen knowledge and skills, by enhancing the overall learning experience.”
Leanne Everett, Principal, Phoenix Academy
What worked well
- Students had a safe, low-stakes space to practise, try again and build confidence over time
- They could work through more scenarios in less time, helping them strengthen critical thinking, communication, active listening and response skills
- Automated feedback provided immediate, structured guidance on how to improve
- The approach worked especially well for beginner students and those building language confidence
- Significantly reduced time required from trainers and instructors to provide students with a similar volume of practice and contextualised feedback.
Challenges faced
- Trainers needed a shared understanding of AI’s role, limitations and ethical use in learning
- Scenarios had to be matched to unit requirements, including performance criteria and knowledge evidence, so students were practising the right behaviours
- Students needed clear instructions and support, especially when moving between existing learning systems and the AI tool
Key insight
- Trainers needed a shared understanding of AI’s role, limitations and ethical use in learning
- Scenarios had to be matched to unit requirements, including performance criteria and knowledge evidence, so students were practising the right behaviours
- Students needed clear instructions and support, especially when moving between existing learning systems and the AI tool
Next steps
Phoenix Academy and CIT will continue testing and refining the approach across other programs.
They will explore how AI chatbots can support more advanced, customised scenarios, giving students wider exposure to real workplace situations and stronger preparation for assessment.