Cantina Labs
Designing for creator trust
The hardest part of the Cantina pivot wasn’t the technology itself — it was trust. We were asking people to let an AI system represent their voice, likeness, and identity, which meant the experience had to feel credible from the very first interaction.
If the product felt confusing, overly automated, or emotionally off, users would hesitate. That meant the design had to do more than look good — it had to create confidence.
Moving beyond novelty
Cantina also needed to evolve from an experimental social product into something creators could take seriously. That created a tension in the experience: how do you keep the product approachable and exciting while also making it feel professional, controlled, and capable?
The challenge was to shift perception without losing the personality of the product. We needed the interface to signal that this was no longer just a prototype or a playful consumer app — it was becoming a real creation tool.
Balancing mobile and desktop
Another challenge was that the product had to work across very different surfaces. Mobile was useful for access and convenience, but desktop was where the deeper creation work needed to happen.
That meant I had to think carefully about how the experience translated across devices. The product needed to feel connected, but not identical. Each surface had to play a specific role in the workflow without breaking the overall system.
Making the AI feel understandable
AI products can easily feel opaque, especially when the system is handling something personal. In this case, the product had to communicate what the AI was doing, what the user was approving, and where they still had control.
The challenge wasn’t just to create a better interface — it was to make the experience feel transparent enough that people would trust it with something deeply personal.
Why this mattered
This was not a standard UI problem. It was a product and psychological problem at the same time. If we couldn’t solve trust, continuity, and clarity, the rest of the experience wouldn’t matter.
That’s why the challenge section matters so much: it shows that the work was not just about execution, but about understanding the deeper product tension we were trying to solve.
Strategy: Design for trust first
Once we understood that trust was the main issue, I focused the experience on making the product feel clear and controlled. Users needed to know what the AI was doing, what they were agreeing to, and what they could still change. The goal was to remove uncertainty wherever possible, especially in places where the product was asking for personal input.
Trust doesn’t come from magic, it comes from clarity. Instead of treating the product like a black box, I designed it as a transparent system where users understood what was happening, what they were approving, and what control they retained at every step.
Make desktop the serious workspace
Mobile was important for access, but desktop was where the real work needed to happen. I treated desktop as the place for deeper creation, more detailed controls, and better fine-tuning. That let us give creators a space that felt more capable without forcing too much complexity onto mobile.
Keep the experience connected
The product needed to feel like one system even though it worked across multiple devices. I designed the flow so users could start on mobile and continue on desktop without losing their place or momentum. That continuity mattered because the process of creating and refining a digital likeness takes time, and the product should help users stay in flow instead of making them restart.
Build around clear handoffs
Rather than making each screen do everything, I broke the experience into steps that made sense for the user’s intent. Mobile handled lighter, faster tasks, while desktop supported the more detailed parts of the workflow. That division helped keep the product simpler to understand while still supporting a more advanced creation process.
Why this strategy worked
This approach gave the team a clear direction: make the product easier to trust, easier to use, and better suited for serious creation. It also gave me a framework for the rest of the design work, because every decision could be measured against the same question: does this help the user feel more in control?
Execution: Leading the desktop experience
I moved into the desktop team with a specific goal: make Cantina feel like a professional product. That meant designing for more advanced workflows, better control, and a clearer path from setup to refinement. Desktop became the place where users could do the heavier work without the experience feeling overwhelming.
Extending the design system
To support that shift, I expanded the design system so the product could stay consistent across mobile and desktop while still serving each platform well. The challenge was not just making the UI match, but making sure the behavior made sense in both contexts. I wanted creators to feel like they were using one connected product, not two separate ones.
Designing the handoff between devices
I also designed the transition between mobile and desktop so users could begin on one device and continue on another without losing progress. That was especially important for a workflow that involved capturing voice, reviewing output, and fine-tuning details over time. The handoff had to feel smooth, or users would drop out of the process.
Using code to move faster
To speed up iteration, I prototyped desktop interactions in React with Claude support. That let me test ideas faster and get closer to production behavior earlier in the process. It also made design and engineering collaboration easier because we could react to working interfaces instead of only talking through static mockups.
Why this execution mattered
This part of the work turned the strategy into something real. The product became easier to use, easier to refine, and better suited for serious creation. More importantly, the team could move faster without sacrificing quality, which helped us keep momentum as the product evolved.
Outcome: A stronger product foundation
The work helped move Cantina from a social experiment into a more serious creation product. By making the desktop experience more capable and the flow more coherent across devices, the product felt more useful for people who wanted to build and refine something real.
Better creator control
A big part of the outcome was giving users more confidence in the process. The product felt clearer, more manageable, and more suited to the kind of personal input required for AI cloning. That mattered because trust wasn’t just a design goal — it was a requirement for the product to work at all.
Faster iteration
The faster design-to-production workflow made it easier to test ideas and keep improving the product without slowing the team down. It also helped engineering spend more time on the core system while I owned a lot of the interface decisions and interaction polish.
What changed
In practical terms, the product became more polished, more usable, and better prepared for serious creators. In strategic terms, it started to feel less like a novelty and more like a platform with a real use case.
What I Learned
"Trust matters more than novelty"
This project reinforced that people won’t use AI meaningfully unless they trust it. A product can be impressive technically, but if the experience feels unclear or out of the user’s control, it breaks down fast. Designing for trust became just as important as designing for capability.
