From HCI to Using AI in Design: A Conversation with Product Designer Bill Guo
Mother of Success · September 26
About the Guest
Bill Guo is a CMU-trained human–computer interaction (HCI) designer who bridges design, engineering, learning science, and psychology.
He serves as Design Lead at an education-technology nonprofit, where he was the first full-time hire, grew the design org from 0 → 2 → 15 part-time designers, and later right-sized to 9 as the team integrated AI practices.
He has also volunteered as a founding designer with early-stage startups in social media, healthcare, and personal care, and loves empowering non-designers with design thinking and rapid prototyping.
“I really enjoy being the bridge—connecting different teams to deliver something innovative, usable, and meaningful. And I’ve learned to be creative and productive within frugality: tight budgets, tighter timelines.”
Beginnings and Background
Xiao He: Bill, thank you so much for joining us at Mother of Success. We’ve collaborated on AI startup projects before, so I know you’re not just a strong designer—you’ve got product and business sense too. Could you introduce yourself to our readers?
Bill Guo: Totally. My background is in human–computer interaction—basically, how people and technology work together. I love being the bridge person who connects teams to ship something useful and meaningful.
Right now I’m the design lead at an educational technology nonprofit. I was the first full-time hire, grew our design team from zero to about two, then to fifteen part-time design members, and now we’ve scaled back to nine, partly because we integrated AI practices into our workflow.
Outside my day job, I’ve volunteered as a founding designer with a handful of “Mother of Success–type” startups—social media apps, healthcare, personal care products. I like working with founders and non-designers to empower them with design thinking, especially prototyping.
A big theme for me is getting the job done within constraints—budget, timeline, people. That’s where AI has been very practical: it helps us stay creative and productive within frugality.
Discovering HCI at CMU
Xiao: Let’s go back to your CMU days. How did HCI shape you, and how did you get into it?
Bill: CMU HCI sits at the intersection of computer science, design, learning science, and psychology. CS helps me collaborate with developers; learning science and psych help me work with researchers and PMs; design is the craft.
Earlier, I started in CMU’s BXA program, combining architecture and computer science. During COVID I realized I craved a community focused on the design–tech intersection. HCI had that—so I pivoted. It trained me to speak multiple “languages” across disciplines and be that bridge.
Designing for Healthcare: The UPMC Capstone
Xiao: You led a capstone with UPMC around cancer care. What was the project and what did you learn?
Bill: Our sponsor was a cancer surgeon focused on a particular bone cancer. The problem: post-treatment monitoring. After surgery, how can providers track recovery and decide who needs to come back in?
We explored smartwatches as sensors—monitoring human activity as a proxy for recovery. Our five-person HCI team handled it end-to-end:
Stakeholder research with surgeons, nurses, and patient-care providers to understand workflows and decision points.
Interface design tailored to surgeons and to patient assistants (two different user groups, two different information needs).
A working prototype that gathered wearable data, was feasible to develop, and was handed over to the sponsor.
Takeaway: Even if you don’t know the domain on day one, structured research + prototyping can get you from zero to credible quickly.
Learning Fast in New Domains
Xiao: Many founders enter unfamiliar domains—PR, social search, hospitals. How should they break in and do credible user research fast?
Bill: First ask: Is it worth the effort to learn this domain? There’s always an opportunity cost. If yes, then approach it like a designer:
Map stakeholders & power dynamics. Who’s involved? How do they exchange value?
Build empathy in actionable ways. Not just vibes—visualize the system, identify workflows, constraints, and handoffs.
Use rigorous methods. There’s a reason trained HCI/psych researchers run studies in big tech.
The #1 mistake I see: leading questions.
“If we built this, would you buy it?”
“How much would you pay?”
You rarely get reliable answers. Before every session, write a short learning objective and gut-check each question: Will this elicit reliable feedback?
If unsure, even ask an LLM to critique your protocol.
Reliable Research and Rapid Prototyping
Xiao: Founders love The Mom Test: delay talking about your product; understand the customer’s world first.
Bill: Agreed. I’d add a practical resource: there’s a handbook with ~100 user research methods—a great starting list to learn study types.
Most bad research comes from under-preparation. Create a simple study protocol with (1) objectives and (2) method choices. Sometimes you shouldn’t bring up your product at all—if it doesn’t serve the objective.
Two methods I rely on:
Directed Storytelling. Ground the conversation in a real, recent experience.
“Tell me about the last time you went to the airport.”
“Tell me about your worst time at the airport.”
Tethering to reality increases reliability.Artifact-Based Research. A lighter cousin of concept testing: put a tangible artifact in front of people—a clickable prototype, a flow, even a sketch—and ask questions about that thing. Tangibility focuses feedback.
Prototyping with AI Tools
Xiao: Founders often over-invest in code too early. What’s your go-to stack for hyper-realistic prototypes on a budget?
Bill: I frame work in the Double Diamond: Discover → Define → Develop → Deliver. Prototyping lives in Develop.
Two core tools:
V0 — great for realistic front-ends with a consistent design system.
Cursor — excellent for “vibecoding,” i.e., iterating with an AI pair programmer to stand up functional UI fast.
Example: During our in-person session in San Francisco, I took a set of user stories and in ~3 hours produced a V0 prototype that looked and behaved like the real product—no back end, deterministic paths. It was enough to get reliable stakeholder feedback—instead of spending three months building an actual app.
Tips:
Write a brief PRD first—user stories, success metrics, scope.
Pick one design system and stick to it.
Use Cursor to match that system’s components.
Add a lightweight back end with Supabase if persistence matters.
Hiring Designers for Different Stages
Xiao: When should a startup bring in a designer—and which type?
Bill: Map your needs to the Double Diamond:
If you need a generalist, test across all four phases.
“You can even test for Discover strength by asking: ‘Tell me about your favorite user-research method and why.’”
Recommended Resources
Bill: Acquired (podcast). It gave me the language and mindset of business, complementing my design and CS training. Designers should learn how business works; this is a practical way to start—during your commute or morning run.
Bonus: How AI Is Reshaping Design Teams
Bill: Design is a team sport. With AI, everyone’s workflow changes—so team composition shifts too. Designers now enter earlier in the process; devs design better UIs; PMs build prototypes.
Team impact:
Smaller, more generalist teams
Faster cycles
Higher expectations
Better validation and transparency
Xiao: Early teams often hire engineers first. But bringing design earlier—from user interviews to prototypes—saves engineering time and reduces miscommunication.
Bill: Founder wisdom is using design thinking as a resource allocation framework: hire the right skill for the right phase so teams move faster with fewer mis-hires.
Tools, Methods, and Frameworks
Double Diamond — Discover → Define → Develop → Deliver
Directed Storytelling — Grounded, real interviews
Artifact-Based Research — Tangible concept testing
V0 / Cursor / Supabase — Prototyping tools
User-Research Methods Handbook — ~100 methods
Acquired — Podcast on business fluency
Additional Notes & Resources
William Lidwell, The Pocket Universal Principles of Design: 150 Essential Tools
Amazon linkThe Double Diamond Framework — Design Council
Current Practice: “New Wine in Old Bottles”
Why This Matters:
If you’re a designer wondering what skills to build next—or a dev/PM/marketer curious how your design collaborator thinks about expanding capabilities—this breakdown shows what’s actually working in practice right now.
Hook:
“I’m putting new wine in old bottles—not fancy, but it works.”
Double Diamond + AI Tools in Practice
Discover — Figure Out the Problem
NotebookLM: Learn about new projects (at the gym, on commutes); get new team members up to speed.
Zoom / Otter: Smarter note-taking and recording.
Define — Nail Down What We’re Solving
Claude as a thinking partner: “You’re a critical consultant. Challenge my thinking, don’t flatter it.”
Miro / FigJam: The usual synthesis tools—now with AI clustering.
Notion AI: Cleans up messy docs and notes.
Develop — Build and Test Ideas
V0 / Cursor:
Example: At RegHero, we spent 3 hours prototyping what we planned to build over 3 months.Not saying we should build it yet—but we could put something real in front of users fast.
Pro tip: Write a PRD first—user story, success, in/out of scope.
Pro tip: Pick one design system (e.g., shadcn)—avoid random button variations.
V0 integrations: Build back-end prototypes without bothering engineers—great for user testing.
ElevenLabs: Voice prototypes.
Google Colab: Train a “good-enough” AI model to demo a feature yourself.
Deliver — Ship the Thing
Weavy: High-fidelity execution.
Runway, Sora, Keling AI: Video/animation tools for production-level assets.