Figma, Axure, Lucidchart. Voiceflow, Onereach, Mix, Copilot. Tools are just vocabulary and I've always been a fast reader.

Five years. Started small: a startup, a blank canvas, a lot of figuring things out. Now I’ve designed conversations for some of the world's largest companies. The space grew, and so did I.

Mistral, Claude, OpenAI. Across each, a handful of models. Different voices, different strengths, each one teaching me something new.

Two years. It started with a single response from a single model, a small experiment with a borrowed team. Most recently, I transformed a legacy IVR into an agentic voice agent. The models got smarter and so did the work.

It starts with questions - requirements, limitations, what business needs and what engineering can bear. Then I sketch, mapping the invisible logic and reasoning that lives behind a conversation before a single word is written. From there I design, whether that means a decision tree or a carefully engineered prompt, depending on what the project calls for. I never work in a vacuum; I loop in business, product, design, and engineering throughout, and I rely on a trusted set of eyes to catch what I've missed. The process doesn't end at launch. I follow the conversation into production, watching the data to find where people get lost, and I use that to make it better.

It started with people. In UX I learned to obsess over how someone feels when they interact with a product - every click, every moment of confusion, every small win. Then AI entered the picture and something clicked. Here was a medium where the interface wasn't a button or a screen; it was language itself. The most natural thing humans do. I wanted to be in that space, making something that didn't just work, but felt effortless for the person on the other end. That mission hasn't changed.