What We’re Seeing
Most AI products today are
half-built, not half-finished.
Founders are shipping faster than ever. AI tools make it possible to go from idea to working prototype in days. But speed without design thinking produces products with two consistent problems.
Completeness
The core flow works. But edge cases break. Failure states are missing. Users hit dead ends. The experience falls apart outside the demo.
Taste
The product functions. But it has no personality. No design sensibility. It feels assembled, not crafted. Users can tell, even if they can’t say why.
Trust
The AI works. But users don’t trust its output. They can’t tell why it made a decision, what to do when it’s wrong, or whether to rely on it at all.
Direction
The team knows what they’ve built. They don’t know what to build next. There’s no framework for prioritising — just pressure to ship more.
Why This is Different
Others just ask what to build; we ask harder questions.
What should the AI do and what should it leave to the user?
What happens when the output is wrong?
How does a user build trust with a system they can’t fully predict?
What does the experience feel like when the model is uncertain?
What does the experience feel like when the model is uncertain?
Our Process
Inside the two weeks.
We move fast. AI synthesises user research, competitive scans, and market signals in hours, not weeks. We align on your business requirements, define the problem statement, map your users, and agree on the use cases worth designing for. What we produce:
We pick two to three major use cases and design them end to end. For each one: the happy path, the failure states, the AI behaviour, and what the user needs to understand at every step. We’re not wireframing screens, we’re designing the full experience of working with AI.Clients often come to this stage with partially built prototypes as a starting point. We meet you where you are. What we Produce:
We validate. If a design isn’t solving the problem or meeting your requirements, we go back. This isn’t a linear handoff, it’s a quality gate. We resolve edge cases, close the gaps, and make the prototype production-ready. What we Produce:
You get a final front-end code prototype, not a Figma file with a design system and reusable components already set up in the codebase. Plus agent-friendly documentation your dev team can actually build from. What we produce:
Deliverables
What you’ll walk away with.
Interactive code prototype
A working front-end prototype your dev team can build directly from.
AI behaviour map
For each use case: what the AI does, where it can fail, what the user needs to know.
Failure state design
Every place the experience can break, designed — not left to engineering to figure out.
User flows & journeys
Mapped against your real users, not assumed personas.
Design system foundations
Reusable components set up in the codebase from day one.
Agent-friendly documentation
Project documentation structured for modern AI-assisted development.
Roadmap & direction
A clear, prioritised path forward so the next step is never a debate.
Two types of founders
come to us for discovery.
AI You have a working model and a product idea. Maybe you’ve already built something. But the experience isn’t there yet — it lacks completeness, trust, or taste. You need a UX team that understands AI products, not one that has to be taught what a language model is.
You’re under pressure to ship AI features. You know what the AI can do technically. You don’t yet know how users will interact with it, trust it, or abandon it. Discovery gives you that clarity before dev starts — not after.
Your questions, answered.
Yes, and this is increasingly common. Many founders come to us with something already built using AI tools. Discovery in this context means evaluating what exists, identifying completeness gaps, and redesigning with taste and trust built in.
We deliver a front-end code prototype, not a Figma file. Components and design system foundations are already set up in the codebase so your dev team can build directly from it.
Yes. Understanding what your AI can and can’t do is part of how we define use cases. We work alongside your technical team to map model capabilities and design around real constraints, not assumed ones.
Two weeks, typically. We cover two to three major use cases end to end, happy paths, failure states, and AI behaviour, with an interactive prototype and direction for the full product at the end.
Discovery feeds directly into the build phase. If you’re continuing with us, your designer is already one sprint ahead of development. If you’re taking the work in-house, your dev team has everything they need to build without ambiguity.
Not sure where your
product stands?
Start with a conversation. We’ll tell you within 24 hours
whether discovery is the right next step and what it would
look like for your product specifically.
Connect with us
Let’s talk about your product.
Hate contact forms? Direct Contact!