What does it really take to move from a promising AI idea to a learning experience that actually works?

In this session, the Pukeko team shares the behind-the-scenes story of developing AI role-play agents for eLearning, focusing not on the final product, but on the rigorous testing, feedback and iteration that shaped it.

Starting with a clear need to create more realistic practice for human skills, the team at Pukeko Learning Solutions explored how AI could simulate meaningful learner interactions. Early prototypes quickly revealed gaps: inconsistent responses, unclear feedback and experiences that did not always feel authentic or useful.

Building on our other online events in the National Webinar Series on 2 and 23 September, this session explores the messy middle. You will see how different testing approaches were used to evaluate the AI’s performance, how stakeholder and learner feedback exposed unexpected issues and how iterative design cycles refined both the technology and the learning experience.

You will hear what worked, what did not and what had to be rethought along the way, including the challenge of balancing technical capability with learning design intent.

Rather than a polished showcase, this is an honest reflection on the process of building, testing and improving AI-driven learning. You will leave with practical insights into how to approach experimentation with AI in your own context and design feedback loops that lead to better, more human learning experiences.

Speaker:  Cain Prentice, Pukeko Learning Solutions

Cain Prentice helps organisations improve performance through learning that works. With international experience across finance, government, not-for-profit and corporate sectors, he has seen plenty of dull, box-ticking training programs and dislikes them as much as the people doing them. His mission is to turn training from a chore into engaging learning that makes a real difference on the job.

At Pukeko Learning Solutions, Cain works with organisations across Australia and New Zealand to design practical, engaging learning experiences built around stories, scenarios and practice. His recent work explores how AI can support more realistic role-play, personalised feedback and stronger opportunities for rehearsal in workplace learning, helping people build capability through practice and show up more confidently in real workplace moments.

We invest significant time and energy into learning—yet much of it never makes it into day-to-day behaviour.

The real challenge isn’t access to knowledge, but what it takes for learning to hold up in the flow of work. In this session, we share a real-world case study from Orion Group’s design and application of Leadership OS—a digital learning system built to support ongoing practice, not just insight.

We focus on a small set of design decisions that shaped what actually happened in practice. We’ll explore how elements like weekly rhythm, participant choice, and micro-practices influenced engagement over time—and where they created friction instead. This includes where flexibility became overwhelming, where digital support needed stronger human reinforcement, and what we had to rethink to sustain participation.

This is a practical, in-progress story rather than a polished success. The focus is on what shifted our thinking about designing for behaviour change, not just knowledge gain. Attendees will leave with concrete design considerations they can apply to any learning context where the goal is not just to inform—but to change what people do.

Speaker:  Chris Rosato, Orion Group

Chris Rosato is Founder and Chief Learning Officer at Orion Group. His work spans leadership development, digital product thinking, and organisational transformation. He focuses on helping people build the capabilities needed in complexity—judgement, adaptability, and the ability to respond in real time. Chris has been central to the design of Leadership OS and is interested in how learning systems can create stronger signals of growth, support consistent practice, and translate into meaningful change in day-to-day work.

Anti‑competitive behaviour is a high‑risk area for NZ Post, where many of our partners, suppliers, and even customers are also our competitors. This creates situations where everyday interactions have potential to slip into cartel‑type conduct, carrying severe legal and reputational consequences.

To respond to this, we designed a 15‑minute digital learning module that could be deployed enterprise‑wide and shared further into the NZ Post group. It targeted a broad, varied audience including managers, specialists, and key operational groups.

The solution needed to feel relevant to everyone while still addressing the needs of a critical, high‑risk cohort that required tighter tracking and reporting. Our learning plan structure allowed the module to be assigned to mandatory audiences, with detailed reporting across tiered risk groups.

The content was intentionally crafted to navigate the shades of grey in competition‑law conversations – from pricing discussions with aggregators, to customer allocation scenarios, to casual chats at BBQs. Realistic NZ Post scenarios helped learners recognise how quickly they could move from acceptable behaviour into red‑zone risk, even with good intentions.

We will walk through how we used performance needs analysis, scenario‑based learning, and human‑centred design to turn complex legal expectations into a practical, relatable digital experience. We’ll also share how partnership with Legal & Regulatory ensured confidence in both the accuracy and sensitivity of the content, while the digital format enabled reach, consistency, and actionable insights through detailed reporting.

Speaker:  Izette du Toit, NZ Post

Izette du Toit is the Digital Learning Manager at NZ Post and a leader at the forefront of digital learning innovation. She specialises in developing, implementing, and overseeing learning solutions that are modern, scalable, and grounded in real organisational challenges. Izette is known for pushing the boundaries of what digital learning can do, exploring new technologies and methodologies to create experiences that genuinely support performance. She brings strong capability in learning strategy, programme leadership, and continuous improvement, with a focus on solutions that are intuitive and impactful. Izette works collaboratively with teams across NZ Post to solve complex problems and deliver digital learning that is future focused, accessible, and aligned to business priorities.

The Benefit Sharing Compassis an online, narrative-driven serious game developed by Geo AR in co-design with Te Kotahi Research Institute, University of Waikato. The project addresses a critical gap in implementing international agreements for fair and equitable benefit-sharing derived from access to genetic resources, which hinders economic and social outcomes for Indigenous peoples and local communities (IPLC). The game’s target audience includes researchers, tertiary students, and new product developers who need a practice-based, accessible tool to build confidence in navigating the benefit-sharing process.

The solution immerses players in three complex, multi-level scenarios focused on Digital Sequence Information (DSI). Players assume dual roles, starting as a research scientist negotiating consent with indigenous people and allocating budgets for non-monetary benefits, and later shifting to a product development lead managing commercialization and regulatory compliance.

Instructional design aligns with adult learning principles, utilizing Kolb's experiential learning and situated learning theories through active, scenario-based gameplay. A core mechanism of the game provides cumulative feedback, scoring player decisions on Fairness, Equity, and Relationship-building. Performance is rewarded with a Bronze, Silver, or Gold Tiki, which visually reinforces the impact of choices on achieving a "cultural license to operate" with IPLC. This structure aims to shift the learner's approach from compliance-based to values-based thinking.

The game is set up for multilingual access, answering significant interest from foreign research institutes. Learners also have access to an extensive Resource kete hosted by the University.

Speaker:  Katie Riddle, University of Waikato

Katie is a Research Officer, working at the interface of Mātauranga Māori, Indigenous Cultural and Intellectual Property (ICIP), Access and Benefit Sharing (ABS), and Digital Sequence Information (DSI) at Te Kotahi Research Institute.

She has worked on the MBIE Endeavour Tikanga in Technology Project, the Science for Technological Innovation (SfTI) Spearhead IP-Related Impact Activities Project, the SfTI Veracity Technology Spearhead Research Project, the Tauhokohoko Indigenous Trade Project, and most recently, the MBIE Endeavour Fund Fish Cell Production Systems (FCPS) for Sustainable Seafood and Marine Innovation Project with Plant & Food Research. She also serves as a Co-Lead of the International Indigenous Forum on Biodiversity Working Group on DSI at the United Nations Convention on Biological Diversity.

New Zealand's pastoral farming sector is struggling to attract and retain the next generation of workers and traditional training hasn't kept pace.

This session tells the honest story of how we tackled that gap by building a digital learning programme from scratch for entry-level farm workers aged 16 and up.

We'll walk you through the real design challenge: creating competency-based digital learning for practical, visual learners who'd rather be outside than in front of a screen. That meant rethinking everything from how we wrote learning objectives to how we designed assessments that actually prepare young people to demonstrate competence in a real farm environment.

We'll share what worked, what surprised us, and what we'd do differently, including the lessons learned around pitching content at the right level, collaborating with subject matter experts, and making assessment feel like a natural part of learning rather than something to dread.

Speaker:  Milly Bekx, Optimism

To ensure equitable access to high-quality nutrition education nationwide, Food for Thought partnered with Cognition Evolve to create Nutrition Quest – an award-winning digital learning game that transforms a traditional face-to-face programme into an engaging, gamified learning experience.

Designed for students years 5&6, Nutrition Quest uses immersive storytelling and interactive challenges to help learners understand nutrition science and apply it to everyday food choices.

Speaker:  Mindy Wigzell, Foodstuffs

Getting a grant to develop content to be shared across your whole sector is one thing, creating engaging and creative content on a budget when you work for a charity is quite something else.

Join Sam and Saran as they tell you their rags to riches story, describing their experiences developing the flipped learning programme Whiriwhiri - Supported Decision Making for the intellectual disability support sector.

This will include identifying need, finding funding, content creation, technology choices and delivery - this is not a sales pitch, they will talk honestly about the challenges, the personal learning, the mistakes and the fun they have had over the past year.

What is Whiriwhiri - Supported Decision Making? It's a culturally informed approach to decision making that upholds the mana of the person most affected. The flipped learning programme aims to develop the skills of people supporting those with disabilities, particularly intellectual disabilities, who have decisions to make.

Speaker:  Saran Read, Hohepa Hawkes Bay

Saran Read (Ngāti Kahungunu ki te Wairoa, Rongowhakaata). Following on from a family of educators, Saran has been involved with adult learning for 20+ years. Her teaching experience includes ESOL, Adult Literacy and Numeracy, Business/Project Management, Financial Literacy and Workplace Education.

Her most recent roles have focused on programme and curriculum design where she is committed to developing digital and in-person training resources that are engaging and responsive to the diverse needs of learners. With a background in Māori Art, Design and Weaving, she brings creativity, cultural insight and strong sense of place to her work.

She is passionate about shaping learning that not only builds capability but also fosters belonging and respect for the cultures and communities of Aotearoa.

Exploring how we designed a 360 virtual marae experience, navigating cultural responsibility, technical constraints, and learner confidence. An honest look at what worked, what didn’t, and what we learned along the way.

This session offers a candid, behind-the-scenes look at Synapsys’s journey into using AI within learning and development. Rather than presenting a polished success story, we will share the real experience – what prompted us to explore AI, what we actually did, and what happened next.

Our work has spanned both using a wide range of existing AI tools and developing our own tailored models to support learning design. Through this, we’ve tested what’s genuinely useful in practice versus what looks promising but doesn’t hold up in real-world delivery.

This will be an interactive and engaging session, with participants invited to actively explore examples, reflect on decisions, and contribute their own perspectives. Attendees will be asked to weigh in on real scenarios, compare outputs, and discuss trade-offs – mirroring the kinds of decisions learning designers are making right now.

We’ll walk through practical, worked examples across design and delivery, highlighting where AI meaningfully improved speed, quality, or learner experience – and where it introduced new complexity, risk, or rework. Alongside this, we’ll unpack the realities of integrating AI into existing workflows, collaborating with clients at different stages of readiness, and navigating evolving expectations, ethical considerations, and concerns about accuracy and trust.

This is not a showcase of tools or a polished “AI success story.” It’s an honest, participatory exploration of experimentation in practice – what worked, what didn’t, and what we would do differently.

Speaker:  Phil Garing, Synapsys

Phil is the Managing Director of Synapsys NZ www.synapsys.co.nz , with over 30 years of experience solving people and process challenges across the corporate, government and education sectors. His passion is in ensuring that investment is driven out of a clear vision for business benefit, and that it is truly measured for impact. Phil is Tangata Tiriti, and proud of Synapsys' mahi in te ao Māori. He has successfully led and managed multiple projects that involved complex challenges, diverse stakeholders, and sensitive environments. Additionally, he has contributed to the development and governance of several learning-related initiatives and networks, such as Learning City Christchurch, the Workforce Development Council Design Group, and EdTech Executive Council. He is currently the NZATD L&D Strategist of the Year.

In this session, we will share two practical case studies that demonstrate how thoughtful digital design can shift behaviour, scale expertise, and meaningfully support organisational capability.

Reimagine Care Learning (RCL) was a project developed to address the need for faster speed to competency and on-the-job critical thinking for inductions and upskills. The challenge was not just knowledge and skills, but confidence, empathy, and behaviour as well. We created a blended digital experience that was human centred, and combined storytelling, scenario based practice, and reflective learning. The result was a measurable uplift in customer experience, quality, improved learner engagement, consistency and stronger alignment to organisational values for a diverse and geographically dispersed workforce.

Expert was a project to implement our new AI-based knowledge platform that holds business critical information across various departments. We designed and delivered targeted training for these roles—frontline agents, Escalations, and Support Teams—ensuring a smooth rollout and confident AI adoption. We meticulously validated AI content against existing resources, enabling decision makers to streamline processes making it easier to transition and migrate into the new Expert platform. This reduced reliance on individual SMEs, improved accuracy, and boosted efficiency.

Across both projects, we will share what worked, what didn’t, and the lessons we learned along the way. You will leave with practical, actionable ideas for designing digital learning that genuinely meets business needs.

We will be presenting with Punit Parekh and Bhawna Raina too - I will submit a separate Expression of Interest for them.

Speaker:  Yasmin Kachwala, 2degrees

Yasmin Kachwala is a Learning and Organisational Development (L&OD) Manager leading capability across RCL and Expert, shaping scalable, digitally enabled learning that drives clarity, consistency, and measurable performance outcomes. She specialises in behavioural learning, designing experiences that move beyond knowledge transfer to embed real change in the flow of work. Passionate about digital innovation, Yasmin leverages platforms, data, and modern design to create intuitive, impactful learning at scale. Known for her pragmatic, people-centred approach, she partners with stakeholders to translate complex needs into effective learning strategies.