The art of decision making: prototyping public services in the age of AI

Theme: Prototyping

Prototyping is often treated as a low-fidelity checkbox - a quick mock-up before the "real" work begins. In practice, it's one of our most powerful tools for validation, alignment, and uncovering constraints early.

As AI reshapes how we design and build, it's changing how we think about fidelity, speed, and the lifecycle of a prototype - not just how fast we can make one. Using a five-layer framework (think of it as a cheeseburger), this session breaks down why we prototype, what's in scope, how we choose fidelity and speed, how long a prototype should live, and how it helps us collaborate as a team.

Drawing on craft principles and recent hands-on work with the Department for Education, we'll examine how to treat fidelity not as a deliverable or tool choice, but as a research variable: something to intentionally dial up or down depending on the question you're trying to answer.

Participant takeaways:

• How to treat fidelity as a research variable - choosing the right level of realism to answer your actual research question, rather than following tool-based conventions
• Practical ways AI can accelerate prototyping - from AI-assisted sketching to vibe-coded, end-to-end prototypes built in a single day
• How to rethink prototype lifecycle in an AI-accelerated world - including when to build a "full set" of prototypes, and when it's time to clear the plate and start fresh
• Strategies for using prototypes as a common language to align policy-makers, developers, and researchers - even when your team isn't comfortable editing code