Update your team.
I help teams think like an architect.
AI didn’t change the job — it changed the medium. I work with your designers to design connected systems for humans, not one-off outputs for agents.
Prepare to solve any problem
I’ll send you exactly what your team will learn straight to your inbox.
Cogito, ergo sum.
“I think, therefore I am.”
René Descartes
Design for the agentic era
Agents change what the work is, on two fronts. Internally, in how your team operates and ships. Externally, in the agentic experiences your customers now expect. Most design teams still treat AI as a feature, not as the material they design in, for, with, and against.
That is where I come in. I work with your designers to design for both: the internal workflows that make the team faster, and the customer-facing agentic products they ship. Practical agent patterns, hands-on exercises, and direct feedback from me on the real systems they build.
From Prompt to System
Move your designers past one-off prompts to agentic systems they can design, rebuild, and hand off like any other product.
Real Tools, Real Workflows
Hands-on with the AI tools your team ships with: prompt patterns, agent workflows, and human-in-the-loop checkpoints, applied to real product work.
Internal and External
Design the agentic workflows that speed up the team, and the agentic experiences customers actually use. Both sides, same discipline.
Direct Feedback and Support
Direct critique on the agentic experiences your team is designing, internal and customer-facing, not generic advice. Small cohort means real attention.
A discipline, not a slogan.
What your team takes away isn’t a tool list, it’s a way of working that holds up when the tools change. It comes down to two moves they’ll practice on real product work.
Object-mapping a SaaS domain — objects, their attributes and actions, and how they relate — before a screen or a prompt.
Object-thinking (OOUI)
OOUI (Object-Oriented User Interface) means designing around the domain’s real things, not its tasks or screens. Before screens, before prompts, model the domain as objects: what they are (attributes), what they can do (actions), and how they relate. That object map is the spec the whole build hangs off, and it is also the ontology the AI reasons over, so the model and the interface describe the same world.
The Snowball Sprint
UX for AI in three moves. Storyboard the work as it should flow. Build a Digital Twin so you can feel it before it exists. Score it on a Value Matrix to decide whether it earns the build. Value is proven before a line of production code is written.
Release notes
Ray, versioned.
Updating your team means bringing in the current version of a practice, not a trend. Every version shipped because the last one hit its limits. Here is the build your team would work with, and how it got here.
v3.0current
Systems-first, AI-native
- AI treated as a participant inside a designed system, never as a feature bolted on.
- OOUI object maps as the working spec and the model’s semantic ground truth.
- Snowball Sprints to prove value before code.
- Defensible, auditable decisions on by default.
v2.0
Designer-engineer
- Stopped handing off. Started shipping production code alongside the design.
v1.0
Experience designer
- Years of enterprise UX. Learned the interface is the easy part. The system is the job.
Update your team.
If you are wiring AI into real, high-stakes work, the hard part was never the model. It is the system it runs inside. That is the work. Let’s talk about yours.