TWOS
Rebuilding human connection in an AI-driven, socially fragmented world
Role
Senior Product Designer — Product Strategy, Research, UX, MVP Design & Launch
Context
VC-backed venture | 0 → MVP → Live campus pilot
01. Context & Origin
As AI products and social platforms continue to scale efficiency, discovery, and content consumption, human-to-human connection has quietly become harder, not easier.
“We exist to reconnect humanity, making life happier.”
Lonely moments are not a failure — they are an opportunity for positive, safe, 1:1 human connection that leads to growth and fulfillment.
The investor believed that algorithmic feeds optimize attention, not presence; social apps optimize identity and performance, not conversation.
At this point, Highline Beta was brought in with a clear mandate: Turn this vision into a real, global product — and launch within one year with daily active users.
02. Reframing the Core Question
How do you design a video-based, global 1:1 connection product that people feel safe starting — repeatedly — with strangers?
Before touching UI, we focused on desirability, matching logic, and emotional safety. Early risks were obvious: video amplifies vulnerability, global products amplify cultural mismatch.
03. Research Phase I — Matching & Behavior
I led extensive secondary research to ground matching decisions in behavioral science rather than intuition.
01. Personality vs. Interaction
Social chemistry at zero-acquaintance is a real-time "state" rather than a fixed "trait." Research shows Big Five measures have little correlation with how much people actually enjoy their first 15 minutes together.
02. Meta-Liking over Compatibility
The strongest driver of conversation quality is "Meta-liking"—the extent to which we believe the other person accepts us. Social enjoyment is rooted in feeling well-received rather than finding a "perfect match."
03. Context-Led Homophily
We bridge understanding via shared value-based traits. Music taste, political interests, and ethical values are far more effective at triggering initial positivity than deep-seated personality traits.
04. The Social Pessimism Gap
People systematically overestimate potential awkwardness and underestimate the cognitive rewards of talking to strangers. Product design must focus on lowering this "entry barrier" of perceived risk.
05. Scaffolding for Introverts
Topic prompting and "smart assistance" significantly improve conversation quality for neurodiverse or introverted users, providing them with the necessary "conversational common ground" to feel safe.
04. Research Phase II — Market Immersion
To complement academic research, I conducted deep market immersion: Downloaded and tested 30+ global friendship apps and completed 200+ real video conversations.
This revealed consistent failure patterns: Over-designed profiles create performance pressure, and most products confuse connection frequency with connection quality.
05. Brand Reset — Designing for Trust
Initially focusing on neutral systems for trust, we strategically pivoted to a more vibrant, youthful brand identity to drive higher engagement and faster feedback loops during our pilot phase.
06. Defining the MVP Shape
A single core ritual: A time-boxed, safe, 1:1 video conversation.
Duration
15-minute hard stop
Philosophy
No messaging or friend lists
Entry
Minimal pre-call info
Closure
Post-call reflection
“We are not another messenger.”
A key strategic decision was the deliberate removal of "Adding Friends." We recognized that adding friends transforms a connection into a "contact," creating the pressure of a digital trail. We wanted to protect the "fleeting magic" of the present moment, ensuring TWOS stayed focused on conversation quality rather than contact quantity like Snapchat or Messenger.
07. Validation — Campus Pilot
To accelerate learning and foster rapid iteration, we launched a high-energy campus pilot at York University. This strategic shift to a more youthful brand resonance allowed us to rapidly exceed 100 active users and observe real-world social dynamics in a high-density environment.
Concentrating Liquidity
To solve the "empty bar" problem, we fixed the matching window to a single daily ritual: 9:00 PM. This created a peak density moment that ensured high match rates and a sense of shared community.
Concierge Matching
We built a custom backend system to manage unmatched users. The founding team personally joined video calls with these students, ensuring every user left with a connection while gathering deep, unfiltered feedback.
08. Iteration & Evolution — Feature Lab
01. The Duration Pivot
Initially, we believed 30 minutes was necessary for depth. By monitoring call drop-off rates and sentiment analysis of conversation content, we saw a clear pattern: small talk peaked at 5 minutes, followed by a "social fatigue" dip at 20 minutes if chemistry wasn't instant.
We pivoted to a 15-minute "Goldilocks zone." It lowered the perceived "time risk" for users while remaining long enough for core topics. To balance this, we added the Extend Call toggle—giving users the agency to continue only when the "magical chemistry" was actually present.
02. Conversational Scaffolding
To eliminate the "What do we talk about?" anxiety, we developed Smart Conversation Prompts.
These weren't just random questions; they were psychologically designed ice-breakers (e.g., "If you had to move to a different country tomorrow, where would it be?") that appeared dynamically on-screen. This "scaffolding" allowed even introverted users to find immediate common ground.
03. Incentivizing Positive Connection
We developed a unique feedback loop by quantifying emotional impact.
By tracking a user’s mood state before and after a session, we created a scoring system that rewarded the "partner"—incentivizing them to be more present and empathetic listeners.
This gamification transformed social vulnerability into a rewarding quest for emotional contribution, rather than just empty badges or engagement metrics.
09. Iteration & Evolution — Visual Maturity
As usage increased, we simplified flows and adjusted visual language to better reflect user maturity. Branding evolved alongside behavior — not ahead of it.
10. REFLECTION
“Some products shouldn’t optimize for engagement — they should optimize for relief.”
Designing for human connection required restraint: Less identity. Fewer signals. Stronger boundaries.