Designing for systems that act on their own.

Design Leader · AI UX EngineerAgentic & AI UX · Microsoft

Seattle, WAMHCI+D, Univ. of Washington

10+ years50+ features shipped

A decade designing human-centered, trust-first AI across conversational, voice, and touch interfaces — for non-deterministic systems that don't behave the same way twice. And I build the tooling, not only the mockups.

Selected work — four case studies ↓

Selected work

The Microsoft Perception overview in Microsoft Defender, shown at an angle: agentic coverage with a flow diagram of incidents through to investigation status.

Agentic Security

Microsoft Security · Perception Demoed at Microsoft Ignite 2026

Designing the human-facing layer around agents that act on their own: how a person reads an unattended run, redirects it, undoes it, and stays accountable. Shipped as Microsoft Perception, generally available inside Microsoft Defender since Ignite 2026. Half open — the launch film, the Ignite demo and what shipped are public. The case study behind it is password-protected; ask me and I'll hand the password over.

35% less triage time25% faster handoffAI-composed UI
The Docufy agreements dashboard and the Word add-in analysing a services agreement, shown at an angle.

Docufy

Microsoft · SharePoint Premium Demoed at Microsoft Ignite 2023

An AI contract lifecycle platform built on SharePoint Premium's content AI services — metadata extraction, document classification, clause analysis. Presented at Ignite 2023 as part of SharePoint Premium; my design framework was adopted org-wide.

3–6 months → under 14 days1.5M agreementsframework org-wide
Passbook by Remitly marketing screens: a debit card and account features shown on two phones at an angle.

Passbook by Remitly

Remitly TIME 100 Best Inventions 2020

An AI-assisted, tri-modal banking app — conversational chat, voice, and touch — built for immigrants navigating English-language and literacy barriers.

67% activationup to 4 features/wk+12% CSAT
The Frictionary live run view and report, shown at an angle: an AI agent working through a task with its reasoning visible, and the resulting readiness report.

Frictionary

My own venture · MVP in pilot

A product that runs a real AI agent through real tasks on your site and shows where it fails, and why. Designed, written and built solo — product, browser automation, data model and the deterministic detection layer behind the evidence claim.

$0.0115 per run8.4s runtimebuilt solo · ongoing

There's more — Puente, Learner Management, and Smart RX — plus experience, skillset and additional impact.

Human control

Placed so agents can run more freely, not less

Reversibility

What can be undone decides what can be autonomous

Trust & red-teaming

Evaluating systems that don't repeat

Multimodal / VUI

Chat · voice · touch, one intent model

Agent-as-user

Designing for a visitor that isn't human

Microsoft · 2026—present

Agentic Security

Human control around autonomous security work — playbooks, sessions, legibility and reversibility — designed from research with 7 security personas.

Microsoft Perception · Ignite 2026
Microsoft · 2021—24

Docufy

AI contract lifecycle on SharePoint Premium. Framework adopted org-wide.

3–6 months → under 14 days · Ignite 2023
Remitly · 2019—21

Passbook

Tri-modal banking for immigrants navigating language and literacy barriers.

TIME 100 · 67% activation
My own venture · 2026

Frictionary

See your site the way an AI agent does — runs a real agent through real tasks and reports where it fails, and why. Designed and built solo, now in pilot.

$0.0115/run · 8.4s · MVP in pilot
Built

MCP workflows

Figma ↔ VS Code ↔ Copilot on one live component state.

~25% faster handoff
Built

Synthetic personas

AI-simulated testing of agent override behaviour.

~30% fewer live cycles
Built

Figma → Adaptive Card

AdaptiveLens — OpenAI + Gemini, emitting schema-valid JSON.

Built

Intent coverage model

Mapped what users asked against what the assistant could answer, then instrumented the gap in Mixpanel.

+12% CSAT
Built

Deterministic detection

Friction found by code, not a model — so no finding can be hallucinated. A model only explains what code already proved.

evidence you can check
drag to pan · scroll to zoom · hover to trace
Agent

Live modelPowered by GPT-4o Grounded in Sayena's case studies, not a canned bot. Billed per question — so make them good ones.

›