Leapfrog is user research tooling that helps user researcher and user research teams distill insights from qualitatitive data, like user interviews. By combining the grounded theory method with artificial intelligence, users can analyze their data at mass.
This project started just before the big AI boom, when I was frustrated with the way we had to manually transcribe and code user interviews. Often times I would find myself dragging notes on online whiteboards, finding data across laptops and losing track of "who said what"? Our interview data would just become too much, and we would have too much data to humanly cross reference. I figured that the way we were doing things could easily be done by a computer.
The start of the journey
I started by writing a post online, and quickly found traction in the form of 700 likeminded researchers, who signed up for my email list.
With that in mind, I started building Leapfrog. I wanted to create a tool that would help user researchers and user research teams distill insights from qualitatitive data, like user interviews. By combining the grounded theory method with artificial intelligence, users can analyze their data at mass.
The idea
User research outputs lots of qualitative data, like recordings of conversations with participants. We often would take turns taking notes, as transcripts were not up to par (although they have been getting better and better ever since).
With new technology, such as LLMs, we could now take some of that data, and create embeddings. This would allow us to do all kinds of analyses on the data pile, bringing user research from the qualitative to the quantitative realm. We would be able to cluster and group data, and even find relationships between different data points. This was what were mainly doing manually.
The tool
Since conversations are semi-structured, we had to rely on a somewhat academic approach to keep our data segmented and organized. It is not uncommon to speak about the same subject throughout a conversation, just like in user research, where participants will often have participants coming back to earlier points. To tackle the irregularities in structure, Leapfrog uses the grounded theory method to highlight and tag data points. This way we take interesting remarks, and remove a lot of the noise from our transcript data.
These tags and highlights can now be used to start analyzing our interviews. We can bring the data into a canvas, and start visualizing it. We can see how different data points are related to each other, and we can start to see patterns emerge. This way the user researcher can focus on the actual theory building, and not the data wrangling.
The product
Intially this project started with building the wrong thing. You might think it is an easy job to build an idea into a product, but in this case we were so deep into AI related interactions that it was hard to prototype any product. I had to rely on good old programming to be able to validate my ideas with the interested parties.
Lessons learned
Building Leapfrog has been a journey of constant iteration. Early prototypes were rough, but they allowed me to test the core value: could AI actually help researchers synthesize qualitative data faster, without losing nuance? The answer was yes—but only when the workflows were designed for real research habits, not just for the sake of using AI.
I learned that user researchers care deeply about transparency and control. Black-box AI outputs are not enough; people want to see how insights are generated, trace them back to the original data, and adjust the process as needed. This led to features like traceable highlights, collaborative tagging, and the ability to audit every AI-generated suggestion.
Where Leapfrog is now
Leapfrog is now a collaborative platform for research teams to upload, transcribe, tag, and synthesize interviews—together and in real time. The tool supports:
What’s next
Leapfrog is in active development, with new features shipping every month. The focus is on making qualitative research faster, more transparent, and more collaborative—without sacrificing rigor. Upcoming work includes:
If you’re interested in trying Leapfrog or want to collaborate, visit leapfrogapp.com or reach out directly. I’m always looking for feedback from real researchers to make the tool even better.
References: Leapfrog website