DESIGNING AGENTIC EXPERIENCE

ABOUT THIS COURSE

Software is no longer something we operate. It is becoming a set of agents that pursue goals, take initiative, and negotiate outcomes on our behalf — often without us watching. The design community is still catching up. Screens, flows, and clicks, the vocabulary of the last two decades do not describe what it feels like to collaborate with, defer to, or push back against a system that has agency of its own. It’s a technology that still doesn’t have well-defined frameworks for how to harness and channel into meaningful experiences. Yet these are the interactions we are now shipping to millions. We need people who can shape this material, not just style it — and who ask sharp questions about what it takes from the world, not only what it gives.

We are entering an era where software no longer simply responds — it acts. Intelligent agents schedule, negotiate, recommend, and decide on our behalf, operating across services, platforms, and institutions with increasing autonomy. Yet as these systems grow more capable, a critical question emerges: what does it feel like to interact with something that has agency? How do we design those interactions to be meaningful, trustworthy, and worthy of the people — and contexts — they touch?

This course invites participants to explore the emerging design space of agentic experiences — where AI systems don't just present information but pursue goals, take initiative, and communicate intent. We examine how technology shapes the feel of agency: how constraints in language, timing, and behaviour define whether an interaction feels empowering or opaque, and whether the people within it retain genuine understanding and control. Participants will investigate how intelligent agents can work within these constraints to create interactions that are not merely functional, but honest, considered, and resonant.

Participants will combine conceptual frameworks with hands-on prototyping, building and testing agentic systems to explore how autonomy, communication, and trust play out in real interactions.

INNOVATION IMPACT for industry participants

This course equips professionals and organisations with the tools to design and integrate agentic systems as meaningful interactions, rather than purely technical features. By developing a clear mental model for agents and gaining hands-on experience with current prototyping stacks, participants can evaluate, iterate, and communicate solutions effectively within product teams. They will also gain a critical perspective and a practical set of questions to guide decisions before, during, and after launch, addressing key considerations such as human agency, trust, reliability, and resource use.

KEY LEARNINGS

• The different eras of AI and how we got to the agentic paradigm.
• Building intuition for intelligence & behaviour as a material to observe and critique
• Opportunity-finding & framing for agentic systems and workflows.
• Prototyping small, functional agentic tools - using the current generation of AI stacks, without getting stuck in engineering detail.
• Critique and Iterate - Does it work? How does it feel to live with this system? How to create explainable, reliable and trustworthy systems that work for people & their unique contexts?

PRACTICAL METHODS & APPROACHES

• LLM primitives for designers - prompting, tool use, and the "jaggedness" of model capability as a design constraint
• The four surfaces of agentic design - goals, context, constraints, evaluation as the craft vocabulary of the course
• Behaviour observation as design research - structured ways to interrogate what an agent actually does, not what its demo says it does
• Rapid agentic prototyping - working fluently with current stacks (Claude / Claude Code, Cursor, v0, function calling, and lightweight local models where appropriate. 
• Critique through use - turning your own prototype into a probe you test on yourself and your teammates
• Ethical and environmental lensing - human agency, consent, tokens-as-consumption, persistence, and the question of what is worth building at all

INNOVATION IMPACT for industry participants

This course equips professionals and organisations with the tools to design and integrate agentic systems as meaningful interactions, rather than purely technical features. By developing a clear mental model for agents and gaining hands-on experience with current prototyping stacks, participants can evaluate, iterate, and communicate solutions effectively within product teams. They will also gain a critical perspective and a practical set of questions to guide decisions before, during, and after launch, addressing key considerations such as human agency, trust, reliability, and resource use.

KEY LEARNINGS

• The different eras of AI and how we got to the agentic paradigm.

• Building intuition for intelligence & behaviour as a material to observe and critique
• Opportunity-finding & framing  for agentic systems and workflows. 
• Prototyping small, functional agentic tools - using the current generation of AI stacks, without getting stuck in engineering detail.
• Critique and Iterate - Does it work? How does it feel to live with this system? How to create explainable, reliable and trustworthy systems that work for people & their unique contexts?

TARGET AUDIENCE

This course is designed for individuals from design, technology, and industry backgrounds who wish to shape meaningful AI-driven experiences. It brings together designers, engineers, product managers, and strategists interested in developing the craft and judgement required for AI-native work, while exploring the interaction layer of intelligent systems.

The course is especially relevant for those who want to explore the interaction-design layer of AI systems, as well as professionals working with AI who wish to move beyond tutorial-driven approaches towards a more reflective, studio-based practice. It is suitable for designers, product managers, engineers, and strategists aiming to develop a making-first perspective on agentic design before deploying it at scale, as well as anyone interested in questioning not only how to build AI systems, but also why they should be built, for whom, and with what implications.