Design + Design Engineering
Reach - AI User Research Interviews
Backed by a Wharton MBA Venture Fund
Process Highlights
Overview
Overview
Reach is an AI-powered platform for scalable qualitative research. It replaces slow, manual interviews with real-time, adaptive voice conversations - conducted, transcribed, and analyzed entirely by AI.
The system lets researchers upload questions, run interviews at scale, and extract structured insights from an interactive dashboard.
The project was driven by a belief that AI can democratize deep user understanding. Designing and building Reach taught me how to ship fast, validate assumptions, and push the boundaries of what language models can do in real-world contexts.
Timeline
One Month
Team
Designer, Developer
Notes
We were accepted into the Wharton MBA x CCV accelerator
What Reach Does
Imagine a Starbucks executive wants to understand how their University City store is performing beyond the numbers. A customer walks in, gets a prompt on the app, and speaks to an AI interviewer for a free coffee. Minutes later, the executive sees rich, structured feedback, no scheduling, no delays.
Research
Analysis
The Current Situation
User research today is slow, expensive, and hard to scale. Interviews require recruiting, scheduling, transcription, and manual analysis, often taking weeks and costing $20–$50 per participant. Surveys are faster but shallow, failing to capture the nuance of real conversations.
Current Solutions
Most teams rely on static surveys, scripted chatbots, or costly agencies. We found that competitors charge upwards of $20,000 and still require users to find their own participants. Others, lack real-time conversational capabilities, making insights rigid and less actionable.
Our goal is to make it easy to create and conduct interviews
Simplicity was at the core of the product.
Automate qualitative user research
Scale interviews with real-time AI
Replace static surveys with conversation
Deliver insights instantly
Integrate seamlessly into products
Reduce research costs
Current user research tactics
Survey forms: Static but user-friendly and easy
Interviews: Rich in data but expensive time-wise
Market research: Doesn’t capture the full picture
Development
Developing the Website
The Landing Page
Any good product requires a good landing page. We went through many iterations and lessons as to how to convey meaning through a story. This was a valuable experience in building startup marketing content.
We reviewed a lot of industry examples, especially Katie Dill’s work at Stripe - great content on storytelling.
Iterations
We wanted the landing page to convey the meaning of the product quickly through text and visuals.
However, we were worried about the intensity of the interaction designs.
Development
The User Experience of Creating an Interview
How do you normally conduct interviews
Participants receive an invite, answer brief qualification questions, and begin a real-time voice interview with AI. The conversation adapts to their responses, capturing deep insights in minutes. Once the interview ends, they’re automatically rewarded with a gift card or incentive.
Receive Interview Code
Brief Diagnostic Test
Conduct Interview
Receive Compensation
Iterations to the voice assistant
What’s the best way to communicate voice feedback?
The desktop log in experience
Need help?
About Reach
Back
Hey there.
Welcome to Reach. We conduct AI user research interviews through phone calls.
You’re currently signed up for:
Reach Demo - Interview
The Reach Team
5 min
$5.00 call
This is not my study
Continue ->
Interview details
Reach
Hey there. We just need your access code associated with the interview you’re taking to get started.
Need help?
About Reach
Submit ->
Enter your access code
Welcome to Reach. We conduct AI interviews through phone calls.
You’re currently signed up for:
Acne Study - Interview
The University of Pennsylvania
5 min
$5.00 call
Continue ->
This is not my study
Hey there.
We need some preliminary information in order to understand a bit about you for our study.
We will send a code to this number to verify its your phone.
Enter your full name
Enter your phone number
Enter your age
Check Eligibility ->
A little about you.
The mobile experience. All of these had to fit inside of a web browser.
Development
Pivoting to Google Forms
Making it easier to create and share interviews
We realized we needed to make it as easy as possible to create interviews and share them. So we took a new direction, and tried to build an AI chatbot style for Google Forms.
This was our pivot.
Someone else had the same idea too...
Demo: Create and Join an Interview
Development
Conducting the Interview - Demo
Audio Calls
We opted for a light-weight audio visualizer, that used our brand colors. It communicated effectively who was speaking and when.
Demo video
Reduce clicks for the user
The user doesn’t want to spend much time on making an interview. Attention is finite, make it as fast as possible.
Keep it clean
Interview can be overwhelming, so we ensured that everything was as bare-bones as possible, while still having acute visual design.
But an unfeasible technology long-term
We realized that the dynamic voice API would not work for complex prompts. This led us to a TTS solution.
Conclusion
Takeways
Takeaways
Designing with AI means thinking beyond static interfaces, it’s about creating systems that adapt, listen, and respond in real time. This project taught me to consider the entire user journey, from recruitment to incentive, and to reduce friction at every step.
I also learned how to design for both web and mobile, ensuring a seamless experience no matter where users engage.
User feedback
Building Reach taught me the importance of rapid iteration: shipping early, testing often, and learning directly from real users. Each development cycle revealed new edge cases, UX challenges, and technical constraints that helped refine the product.
It reinforced that great products aren’t built all at once, they’re shaped through continuous feedback and adaptation.
