Experiment
Nothing will ever beat talking to human beings. Not even AI. But, lately I’ve been exploring new ways to bring user voices into the room; especially when the room is full of stakeholders. And not just quotes on a slide. I mean really bringing users to life.
But in practice, most product decisions aren’t made in research sessions—they’re made later, by people who only ever see quotes on slides. Tone gets lost. Context evaporates. Empathy fades.
This experiment explores a question I keep running into in real work: Can synthetic avatars (powered by real user verbatim) help preserve the emotional signal of user research without misrepresenting the user behind it? Not to replace research. Not to automate empathy. But to see whether we can bring user voices back into the room in a way that sticks.
Other experiments (all types)
This video is decorative. The video has no sound.
User needs
Stakeholders need to understand users—but more importantly, they need to remember them
In many organizations, decision-makers are several steps removed from the people they design for. Research artifacts get summarized, condensed, and sanitized until what remains is technically accurate but emotionally inert.
The underlying user need here isn’t more data. It’s durable empathy—insights that survive beyond the meeting where they were first presented.
What I tried
Give real users a synthetic voice using their actual verbatim
I used real user verbatim from prior research and mapped each quote to an established persona (for example, “Power Buyer with Trade-in” or “Experienced Dealer Agent”).
That verbatim was then used to generate short avatar-based video clips using:
- HeyGen
- Hedra
- Google Veo 3
Each avatar delivered a single, tightly scoped insight—typically under 20 seconds—designed to be consumed inline within a research artifact or presentation.
- No scripts were invented.
- No sentiment was exaggerated.
- The goal was controlled expression, not performance.
The upside: why I’ll probably keep using it
Emotional resonance
Synthetic faces + voices transform cold research artifacts into something closer to a documentary. It’s harder to ignore a frustrated “user” who looks you in the eye.
Speed and control
Compared to real video editing (which is slow, constrained by consent, and often unusable due to NDAs) avatars offer speed and flexibility. I can compose, iterate, and update insights quickly while keeping everything on-brand and presentation-ready.
Accessibility and consistency
Everyone receives the same message, delivered with the same clarity and pacing. That consistency reduces distraction and helps teams focus on interpretation and implications instead of delivery artifacts.
The pitfalls: why it still gives me pause
The uncanny valley is real
When avatars drift too close to realism, they can become distracting—or even unsettling. This limits how long these clips should be and argues for restraint in emotional expression. Subtlety matters.
User verbatim communicated using HeyGen
User verbatim communicated using Hedra
User verbatim communicated using Google Veo 03
