About
We're a UX research team. We're also an AI application research team.
voidvector runs four practices as one line — AI application, UX design, growth & brand consulting, and marketing growth — because your users never experience the product as four separate departments.
Two research disciplines, one team
voidvector started simply: we were user researchers who watched finished features go unused, then watched the AI wave produce features built for AI's sake. Both failures are the same failure — teams building what's convenient to build instead of what people actually use. So we fused the two disciplines: UX research tells us what's worth building; AI application research tells us what's actually buildable right now.
Four practices, one loop — not four vendors
You could hire four agencies: an AI shop that ships features nobody touches, a design studio that draws flows the tech can't support, a growth consultant optimizing a funnel with a broken experience inside it, a marketing agency driving traffic that never sticks. Everyone delivers; the budget dies in the handoffs. We run AI application, UX design, growth & brand consulting, and marketing growth as one loop — research findings become design decisions, and growth data flows straight back into the product.
Evidence, not taste
Usability session recordings, retention curves, visibility data across search and mainstream AI engines — every recommendation we make has to trace back to something we observed. If we can't show you where a conclusion came from, we shouldn't be charging you for it.
AI has to beat the dumb baseline
Most AI features are demos wearing a product costume. We run model selection and evaluation before anything ships; retrieval Q&A and workflow automation go live only when they beat the dumbest workable baseline. When they don't, we'll tell you to skip them.
How we think
What we value.
- 01
Research first
We read users and numbers before we touch anything.
- 02
Opinionated delivery
You get a verdict, not a menu of options.
- 03
Numbers map to moves
Every action is tied to a metric you actually care about.
- 04
AI where it belongs
In the right place — not every place.
- 05
Handover means you run it
If you still need us to operate it, we are not done.
- 06
Honest about what failed
We tell you what didn't work — and we say it first.
Meet the team behind the work.
- 01Clarence Wu Founder & Research Director
Sets the direction and keeps both research practices honest — about what users actually do, and what AI can actually deliver.
- 02Clarence Wu Head of UX Research
Runs user interviews and usability tests — and the information architecture and interface work that follows — turning what people actually do into decisions teams can act on.
- 03Clarence Wu Head of AI Application Research
Leads model selection and evaluation — retrieval Q&A and workflow automation ship only after they beat their baselines.
- 04Wilbert Hsieh Head of Quantitative Research
Leads quantitative research and analysis — turning "did it work" into something measurable instead of a feeling.
- 05Clarence Wu Growth & Brand Consultant
Works on brand positioning, retention, and LTV — turning one-off wins into a growth loop that compounds.
- 06Clarence Wu Marketing Strategy Lead
Runs SEO, GEO, and content strategy — making brands visible where search and mainstream AI engines answer for them.
Careers
We're hiring.
We're looking for UX researchers, AI application researchers, and growth and marketing strategists who want to work the whole line — from user evidence to shipped product to measurable growth. Full-time, Taipei or remote.
From the team
Insights.
Lessons, frameworks, and honest takes on building AI that earns its place in the product.
-
AI applicationsA self-writing wiki only compounds knowledge if someone can catch the page it got wrong.
-
AI applicationsMetrics before models: three questions to answer before shipping an AI feature.
-
AI applicationsWhen RAG retrieval disappoints, the model is rarely the culprit. Your knowledge base is.
Let's find where intelligence moves your metric.
Tell us what you're building. We'll tell you honestly where intelligence moves the number — and where it doesn't.
Book a call