KIID AI Feedback

UX Research

Marketing Events

Social AI Startup

Design Insights

From the Coffee Shop to the Design Room: UX Research at a 0-1 AI Startup

My Role: Marketing & UX Research Intern — user recruitment, interview design, feedback synthesis, findings presentation
Company: KIID AI — an early-stage social AI startup
Timeline: October 2025 – Present

Context

KIID AI is building something genuinely novel: a social platform where users create an AI avatar of themselves and interact with the AI versions of their friends and strangers.

When I joined, the product was pre-launch. The design team was still actively working out what the app was for — what would make someone open it, come back, and invite a friend. My role was technically marketing, but the most important marketing question and the most important UX question were the same thing: who wants this, and why?

Phase One: User Recruitment

Before any research could happen, we needed users. The product had no existing base— we were starting from zero.

I launched a grassroots recruitment strategy centered on a coffee tab at a high-traffic campus location. The idea was simple: cover someone's coffee in exchange for a few minutes of their time to hear about the app and share their reaction. It was low-friction, naturally conversational, and put us in front of exactly the demographic the product was targeting.

Phase 2: From Recruiter to Researcher

What started as a recruitment role quickly became a research one. Every conversation at that coffee shop was a data point. I started noticing patterns in how people responded and realized I was sitting on generative research that the design team needed to hear.

I formalized the process. I developed structured questionnaires to capture feedback consistently, ran feedback sessions where users could interact with the app concept directly, and conducted informal interviews that let people express discomfort or excitement in their own words.

Three themes surfaced repeatedly:

1. The avatar felt uncanny, not personal.

2. The way the AI spoke felt off.

3. Onboarding was too long.

Findings and Impact

I synthesized these themes into a structured findings presentation and brought them directly to the CEO and design team. The goal wasn't just to report problems — it was to frame each finding in terms of what it meant for the product direction and what a response might look like.

Phase 3: Report and Iterate

The team acted on all three:

New avatar design — a visual redesign that moved away from the uncanny middle ground toward a style users found more approachable and less unsettling
Revised AI language — changes to the tone and phrasing the AI used to feel more natural and personalized
Shortened onboarding — a streamlined flow that got users to the core experience faster

Users I spoke to after the changes described the app as something they could see themselves using — "excited" replaced "creeped out" as the dominant reaction. That shift didn't come from a design intuition. It came from listening.

Three themes surfaced repeatedly:

1. The avatar felt uncanny, not personal.

2. The way the AI spoke felt off.

3. Onboarding was too long.

Findings and Impact

Phase 3: Report and Iterate

The team acted on all three:

New avatar design — a visual redesign that moved away from the uncanny middle ground toward a style users found more approachable and less unsettling
Revised AI language — changes to the tone and phrasing the AI used to feel more natural and personalized
Shortened onboarding — a streamlined flow that got users to the core experience faster

Reflection

This internship taught me what UX research actually looks like outside of a classroom or a formal lab setting — and I loved it. There's no polished research protocol when you're at a startup trying to figure out what you're even building yet. You make do with a coffee tab and good questions, and that kind of connection with users is what I want more of.

The team's decision to act on my findings — and to move me into a dedicated UX role because of it — was one of the most meaningful professional validations I've had. I'm still with KIID AI, now working exclusively in UX, continuing to talk to users and iterate on the design as the product evolves. The use cases are still being figured out, the design is still changing, and that's exactly what makes it exciting.

Like what you see? Let's keep this party going ↓