The Summary
Alexa+ Studio is an agent-first developer workspace for third-party brands building skills on the voice-led Alexa+ platform. Developers describe what they want to build, submit their assets, and the agent reads everything, builds their skill’s flows, and surfaces its reasoning for review. The developer’s job is confirmation, not configuration.
I was the lead AI product designer, responsible for the foundational experience: the conceptual model, the interaction architecture, and the design system that runs across it. I defined the platform architecture, designed the integration experience and simulator, and shipped live AI prototypes that drove cross-org alignment. I reduced third-party developer workflow phases by 45%.
I delivered:
- AI-powered prototypes to showcase interactivity
- Agent personality specs & SDK guardrails to ground agent behaviors and responses in reality
- Design system tokens and components to present a visually cohesive web experience
My primary tools included:
- Figma with MCP (defining the design system components, major page layouts, tokens)
- Cursor (my primary prototyping tool)
- Anthropic API (for data processing; general agent for collaboration and authoring)
The Situation
The existing Alexa Developer Console was built for deterministic voice experiences and logic structures. When Alexa flipped to the probabilistic model, I overhauled the console to help developers quickly integrate, fine-tune, and simulate their customer experience.
Implications from the legacy integration experience included:
- There was no way to evaluate an Alexa integration before committing to a full build.
- It couldn’t surface what an AI agent would build on behalf of the developer
- Developers could not audit, update, or challenge data inferences
- Scaffolded experiences were not feasible
The “before” design you see here combined a step-by-step wizard and jump points to nested functionality:
The Strategy
I focused on an agentic integration experience, enabling developers to provide assets and natural language during onboarding. The agent would audit the information provided in order to infer and scaffold a customer experience. If there were constraint violations between the SDK guardrails and the 3rd party product requirements doc, the agent would flag those for turn-by-turn resolution with the developer.
The tool would quickly move the developer into an experience that combined building and simulation. They would have the ability to preview the customer experience and then dive into the mechanics of how it worked. From there, they could pinpoint parts of the experience that were incorrectly inferred or scaffolded and fine-tune the interaction or data sources.
My primary focus: Zero compounding mistakes
In probabilistic AI systems, small errors amplify downstream. I designed validation checkpoints at every workflow stage so developers caught integration issues before they became embedded in a live skill. Trade-off: more gates to maintain, but errors that are invisible until they ship are worse.
My secondary focus: Autonomy where it’s earned
Not every agent action needed the same oversight. I defined a framework matching oversight requirements to decision weight. Low-stakes actions proceed autonomously, high-stakes ones require confirmation. Trade-off: more nuanced agent logic required, but treating all decisions identically makes the system either too slow or too risky.
The Results
- Reduced workflow phases by 45% by grouping related tasks/goals
- Live prototypes built in Cursor using the Anthropic API and Figma MCP
- Drove cross-org alignment on quality gates and capability decisions
- Established platform architecture, JTBD framework, DXOs, and design tenets