I shape how AI agents communicate.
Whether a user trusts an AI agent depends entirely on how it handles incomplete context, when it chooses to surface what it cannot know, and how it structures its reasoning. I design the systematic response patterns and quality frameworks that make those interactions predictable and defensible.
Featured work.
Authoring the voice and reasoning of an AI agent in a high-stakes domain.
When an agent invents a safety threshold or buries the answer in five paragraphs, the cost is a wrong call on live equipment. The response guidelines I wrote give the agent a consistent voice, a defensible refusal pattern, and a way to reason about its own data before responding.
Read case studyTurning agent quality from opinion into a measurable baseline.
Without a measurable baseline, every change to the agent becomes an argument about whether the response is better. The evaluation framework I built makes that decision evidence-based, with five scoring dimensions, four user contexts, and a control-versus-variant protocol that proves a change is an improvement before it ships.
Read case studyScaling agent behavior without losing quality at the edges.
A single monolithic instruction set works for one product, until rules start conflicting and the agent loads everything for every request. The skills architecture I designed separates the persona from four reusable response patterns, so the agent loads only what the task needs and the content team can refine each pattern independently.
Read case studyTurning content strategy into a workflow the agent can run.
Content strategy work usually sits in a designer's head and a few scattered documents, which makes it hard to share, repeat, or scale. The skills framework I designed turns each step of content strategy into something an AI agent can execute, producing a documented content architecture for a feature before any UI text gets written.
Read case studyAbout.
I'm a senior content designer engineering how AI agents communicate in complex product ecosystems. I build the foundational behavior models that dictate how an agent reasons through data, and the structured skills frameworks that turn content strategy into reusable product infrastructure.