Krissie Shelley — Senior Product Designer & Rapid AI Prototyper
I turn ambiguous emerging technology into clear, testable product experiences — combining deep UX judgment, visual craft, and enough development fluency to build working concepts, not static mockups.
Static mockups can't tell you how an AI product will behave. Working concepts can.
AI products are non-deterministic: answers stream, stall, drift, and sometimes fail confidently. A click-through prototype hides all of that. I build working concepts so teams can test the real behavior — latency, uncertainty, permissions, and recovery — before committing to a roadmap.
I started hand-coding sites in 1998, then spent years designing and engineering complex SaaS for regulated industries. That combination — UX judgment, visual craft, and enough development fluency to ship — is what lets me move from ambiguous idea to testable product in days, not quarters.
Prototypes that run, so the hard questions surface early.
Prototype 01 · Agentic learning
FrizzleNet
A place-based learning agent that turns a family's real surroundings into guided activities — built as a working, deployed prototype rather than a deck.
What it taught me
01 — Uncertainty & fallbacks
Sparse location data and vague requests are normal, not edge cases. The agent offers a safe default activity and asks one clarifying question instead of guessing or stalling.
02 — Human review & permissions
The parent stays in charge. Suggestions are proposals — adults approve, swap, or skip each activity before anything is shown to a child.
03 — Pacing & cognitive load
A live load gauge showed that fewer, calmer steps beat clever ones. Speed of response mattered more than depth of reasoning for this context.
More prototypes in progress — each documented with what it tested, what broke, and what changed.
Designing for the moments the model gets it wrong.
Uncertainty & failure states
I design the path for when the model is unsure, slow, or wrong — so trust survives the bad moments.
- Confidence cues and plain-language caveats
- Graceful fallbacks when data or the model fails
- Clarifying questions instead of confident guesses
- Easy retry, edit, and undo
Permissions & human review
Agents should earn autonomy. I define when the system suggests, when it asks, and when it may act.
- Approval gates before consequential actions
- Tiered permissions for what an agent can touch
- Reviewable diffs and clear audit trails
- Always-visible override and stop controls
Fast experiments are only useful if you know what it takes to ship.
Years of shipping audit, compliance, and analytics platforms — permissions, data governance, edge cases, accessibility — are what keep my rapid AI concepts grounded in how real products behave in production.
A working history across studios, startups, and enterprise.
- 2023 —
Naturally Human Designs — AI Experience Designer & Principal Design Technologist
AI-native studio focused on human-centered AI products, rapid prototyping, and digital transformation.
- Shipped a place-based experiential learning agent on Google ADK + Antigravity (Kaggle Vibe Coding Intensive) with custom prompt frameworks, state-machine execution payloads, and HITL feedback loops.
- Architect multi-step prompt workflows and interactive feedback loops for agentic AI applications, reducing cognitive load and driving engagement.
- Leverage LLM-powered prototyping (Lovable.dev, Cursor, Claude Code, Google ADK) to ship production-ready web applications on compressed timelines.
- Designed secure multi-client digital platforms integrating Resend automation, Supabase backends, and Google Analytics to continuously optimize UX.
- Led divisional rebrand and end-to-end digital transformation for Vet-Ray by Sedecal — admin dashboards, lead-tracking tools, and user behavior flows.
- Facilitate discovery workshops with stakeholders to map complex workflows, validate AI use cases, and align technical capability with real-world needs.
- 2022 — 2023
SalesRabbit — Senior Product Designer
B2B SaaS sales engagement platform serving thousands of field sales reps.
- Bridged product, design, and engineering through a legacy-to-React platform transition with scalable design standards.
- Built high-fidelity prototypes and internal tools that streamlined B2B pipeline management and reduced partner integration friction.
- Ran deep qualitative research with field sales clients, translating friction points into features that improved retention.
- Partnered with product management and engineering leaders to deliver complex, high-priority features aligned to business strategy.
- 2021 — 2022
Freelance UX Design — UX Strategist & Consultant
Partnered with early-stage startups to simplify complex, data-heavy SaaS platforms.
- Overhauled the Blubrry Podcast Services analytics dashboard, turning complex creator metrics into decision-driving visuals.
- Designed streamlined grant-application flows for FarmRaise, simplifying federal compliance forms for farmers.
- 2019 — 2021
Suralink — UX Designer
B2B SaaS audit request management platform for accounting firms and their clients.
- Established the organization's first scalable design system and component library with WCAG accessibility built in.
- Designed secure file-sharing and document-exchange workflows compliant with HIPAA and financial regulatory standards.
- Owned end-to-end design for dashboards tracking multi-stakeholder audit requests, reducing operational friction.
- 2014 — 2019
Nav Technologies — Senior Product Designer
Fintech platform helping small business owners understand and build business credit.
- Designed data-dense dashboards that demystified business credit profiling into clear, actionable steps.
- Shipped iterative improvements to a multi-sided marketplace matching businesses with credit and lending products.
- 1998 — 2014
Various — Web Developer & Designer
Foundational years — hand-coded HTML/CSS websites that shaped a lifelong sense of how code constraints shape UX.
- Evolved from hand-coding into visual design, branding, and UX strategy.
- Built the technical intuition that enables seamless collaboration with AI architects and engineers today.
What I bring into every engagement.
01 — AI Experience Design
- Prompt engineering & workflows
- Human-AI interaction models
- HITL workflows
- Agentic AI UX
- Escalation & override mechanisms
- Feedback loop design
02 — Product Design & Strategy
- End-to-end UX/UI
- User research
- Rapid prototyping
- Journey mapping
- Design systems
- Stakeholder facilitation
03 — Frontend & AI Prototyping
- React
- Tailwind CSS
- Lovable.dev
- Cursor
- Claude Code
- Google ADK
- Supabase
- API integrations
- Semantic HTML5/CSS3
04 — Design & Collaboration Tools
- Figma
- Miro
- Framer
- Adobe Creative Suite
- Affinity
- Jira
- SCRUM
- Git
05 — Accessibility & Compliance
- WCAG accessibility standards
- HIPAA
- Financial regulatory compliance
07 — Let's talk
Let's build
something.
West Valley City, UT · Open to 100% remote