Overview: This regional sandbox targeted the digital transformation of Technical Vocational Education and Training (TVET) by equipping educators and learners with AI-supported pedagogical tools.
- Focus: Ensuring AI-enabled tools meet the diverse needs of the SEA-VET platforms learners across the region.
- Action: Providing VOCTECH with multiple AI chatbot options — varying in multilingual support, contextual knowledge, and personalisation — to encourage internal discussion and align tool selection with student needs.
Context
SEAMEO VOCTECH hosts SEA-VET Learning, a platform designed to play an important role in expanding access to vocational education and training across Southeast Asia. While the platform successfully engages learners in its hybrid trainings, sustaining learner engagement and supporting course completion of its self-paced, asynchronous offerings remains a challenge.
VOCTECH and EdTech Hub identified AI chatbots as a potential EdTech solution to keep learners engaged and motivated in self-paced courses. Although AI chatbots for learning have been used in the basic education space in Southeast Asia, their use in the technical and vocational education (TVET) sector remains largely untested. This sandbox set out to test whether an AI chatbot is a suitable feature for TVET learners, and to guide VOCTECH on whether, where, and how AI can responsibly support self-paced TVET learning across Southeast Asia.
The Challenge
Although AI has become a central focus in education, effective integration requires careful reflection and a clear purpose. Rather than deploying AI for its own sake, institutions need a focused use case that aligns with learners’ needs. In addition, it was critical to the VOCTECH team that platform improvements were aligned with larger strategic goals and sustainable.
The Hypotheses
This hypothesis served as the strategic foundation for our design, positioning AI as a support layer to enable persistence rather than a replacement for human expertise:
IF we enhance course clarity, design quality, and offer structured learning pathways, THEN learners will have clear visibility into learning outcomes and career relevance, and overall user engagement of the learning platform will increase SO that learners successfully complete SEA-VET.net Learning self-paced courses and use the platform to develop job-relevant skills.
Desirability
| IF we integrate an AI Chatbot as on the SEA-VET platform, THEN learners will stay engaged with learning materials, SO the tool must provide support and guidance that learners actually want and value. |
Feasibility
IF VOCTECH decides to integrate an AI chatbot, THEN long-term success will depend on the organisation’s readiness, SO the tool must align with strategic priorities and VOCTECH capabilities.
Viability
IF AI support increases completion rates and platform engagement, THEN the SEA-VET platform becomes a more sustainable regional resource, SO the tool must prioritise learner persistence.
Ethics
IF users are using AI for learning and instructional support, THEN they must be able to trust the information and guidance they receive, SO the design must prioritise accuracy, accessibility and encourage learners to evaluate content, not just consume it.
Validation Process
To ground the project in evidence, we identified critical assumptions during Sprint 0 (Preparation) to determine what needed to be proven true for the AI integration to succeed:
| Assumptions | Outcome & Validation Method | Implication |
| Instructor support is a key engagement in hybrid and online courses. | Survey and interview participants confirmed that instructors and peers are a key support system when learning. 62% of survey respondents said regular feedback or check-ins from instructors or mentors help them stay motivated. | Strategies to increase engagement and motivate learners in self-paced courses should focus on ways to provide support for learners that they would typically receive from an instructor. |
| SEA-VET Learning users would use an AI chatbot if it was integrated into the site. | Based on engagement with our survey and interviews, we determined there was a moderately strong interest in having an AI chatbot on the platform. We also learned that many learners are already using AI when learning online: 75% of survey respondents said they use ChatGPT or a similar tool for quick questions, indicating that AI is a tool people are using when learning. However, we heard some concerns from interviewees about AI hallucinations and over-reliance on AI tools. | Based on the ways people are already reporting that they use AI, a potential use case for SEA-VET Learning could be a chatbot that answers learners’ questions. To differentiate it from existing tools, it could be tailored to the Southeast Asian context and TVET learning specifically. The development of the chatbot needs to be carefully done so it doesn’t provide wrong information and supports learning, rather than providing quick answers. |
| The SEA-VET Learning team will be able to implement and maintain a chatbot on the platform. | Through the workshop, VOCTECH were informed of the costs and other considerations for each AI chatbot option and use case. | VOCTECH has a guide for further discussion to identify which use cases they want to prioritise, what functionality they want to have, and the costs required. |
Sprint Journey
The sprint cycle focused on testing the desirability, feasibility and viability of an AI chatbot to support asynchronous, self-paced learning and facilitate evidence-driven decision-making.
Sprint 1
Testing the desirability of an AI chatbot among SEA-VET Learning users.
Scope
Exploring interest and use cases of an AI chatbot on the SEA-VET Learning platform.
Do
Developed surveys and interview protocols to explore strategies learners use to keep motivated and engaged in self-paced learning, and gauge interest in using AI in self-paced learning.
Measure
Analysed survey responses and interviews to identify trends and themes about self-paced learning and feelings about AI for learning among respondents.
Learn
Used insights to propose AI use cases to the VOCTECH team that had the potential to support SEA-VET Learning users to stay engaged and complete self-paced courses.
Sprint 2
Testing the feasibility and viability of an AI chatbot with VOCTECH.
Scope
Exploring AI options and VOCTECH readiness to integrate AI.
Do
Researched AI chatbot options, considering costs, scalability, and technical specifications.
Measure
Conducted a workshop with VOCTECH, including representatives from the SEA-VET Learning Team, IT department and executive leadership. This workshop aimed to inform VOCTECH of the AI chatbot options and “stress test” the proposed use cases. We aimed to provide a framework for VOCTECH to guide their decision-making, which considers the value of the use case to the learner, the desired functionality, and the costs and capacity needs of AI integration.
Learn
The workshop allowed VOCTECH to discuss which use cases were a priority, what functionality they wanted, and identify where costs or capacity could be a limitation.
Outcome and Reflections
The sandbox confirmed that AI’s primary value in TVET is as a friction-reducer rather than a motivator. Findings showed that learners are already using tools like ChatGPT selectively for quick clarifications and translations to maintain their learning momentum.
The critical insight was the role of AI in providing just-in-time support; learners could use a bot to provide additional explanations or translate unfamiliar terms. This was echoed by a TVET teacher trainer from the Philippines:
I think it’s [AI] helpful also in the learning process. There’s someone that can give you an immediate response when you need [it] most. Because some peers, colleagues are also busy doing some things…but there’s an AI tool that can immediately be accessible in your view.”
This suggests that by solving these immediate “small hurdles,” the AI acts as a safety net that prevents disengagement and potential course abandonment. We reflected that AI’s role is to fill the support gap when human assistance is unavailable, ensuring curiosity or confusion does not lead to disengagement.
Next Steps
These sandbox sprints provided foundational work to assess if an AI chatbot was an appropriate choice to test the hypothesis. These steps were essential to test our critical beliefs and ensure the development of a prototype or pilot was grounded in evidence rather than based on assumptions about what learners wanted or what VOCTECH was able to implement.
Following Sprint 2, VOCTECH’s next steps are to discuss the ethical dimension of design thinking. Given the many debates about responsible AI use for education, it is important to consider whether AI integration is the right choice for the platform. Assuming that VOCTECH wants to pursue AI integration, they must also decide where use cases land on a matrix of impact versus feasibility. This can help identify which use cases could be used as a pilot to further test the impact of an AI chatbot on learners’ engagement and motivation in self-paced online courses.
Related Work
This work is part of the portfolio of projects delivered by the ASEAN-UK SAGE programme. The ASEAN-UK SAGE programme is delivered by the British Council and SEAMEO Secretariat, in partnership with EdTech Hub and Australian Council for Educational Research (ACER), and is an ASEAN cooperation programme funded by the UK.
This sandbox was led by Delanie Honda, Nawaz Aslam and Sangay Thinley.
