This episode is brought to you by EY:
How do you move your AI discussions to AI actions? EY's AI Pioneers podcast miniseries features speakers from pioneering government organizations around the world. You'll hear practical lessons from your peers who have successfully moved from ambition to impact – scaling AI to transform their organizations and how they serve citizens. Explore the series and subscribe here for future episodes on your podcast platform of choice.
Episode Overview
In this episode, we hear from Dan Roelink, Director, Office for AI, Digital.NSW, NSW Department of Customer Service, who explores why the biggest challenge for AI in government is no longer understanding what the technology can do, but reducing the system friction that prevents safe, lawful, and scalable adoption. From an inside-government perspective, the conversation breaks down what it takes to move beyond experimentation and pilots by building the foundations, governance, and operating models needed for sustainable, system-wide deployment.
Key Themes
Dan focuses on turning AI ambition into repeatable delivery by modernising governance, risk management, and assurance. It highlights the core components governments need to scale responsibly, the changing relationship between people and AI as barriers reduce, and the role of partnerships and problem-led collaboration in moving from ideas to production outcomes.
What You’ll Learn
1) Why “Friction” Is the Real Barrier
Why adoption stalls even when the technology is understood, and what types of organisational and system constraints slow progress.
2) Modern Governance, Risk, and Assurance for AI
Why AI requires updated oversight, clearer accountability, and fit-for-purpose assurance approaches to stay lawful and trusted.
3) The Core Building Blocks to Scale
What needs to exist across policy, frameworks, capability, platforms, and operating models to enable consistent delivery across government.
4) What a Low-Friction Path to Production Looks Like
How reducing barriers between people and AI changes the speed and reliability of moving from concept to deployment.
5) The Human and Partnership Dimension
How people, partnerships, and problem-led collaboration create the conditions for scale, not just isolated innovation.
6) Future Scenarios for Mature AI Adoption
What “good” can look like in a mature, low-friction system, and how effective partnerships may need to evolve to support it.
Key Takeaways
- Scaling AI is an operating model challenge as much as a technology challenge
- Governance, risk management, and assurance must modernise to keep adoption safe and lawful
- Real scale depends on foundations: policy, frameworks, capability, platforms, and delivery models
- Reducing friction accelerates the journey from idea to production
- Partnerships and collaboration are essential for sustained, system-wide impact
Why You Should Listen
This episode is valuable for public sector executives, digital and data leaders, AI governance teams, risk and assurance professionals, and delivery leaders working to move AI beyond pilots. It offers a practical lens on what must change across systems, people, and partnerships to enable AI adoption that is scalable, trusted, and sustainable.
Help your peers
Share what you've learned with fellow public servants
