Loopy AI

Inspiration

At a previous startup I interned at to familiarize myself with the user experience of the product

What it does

Session replays are video-like recordings generated for every user interaction with your website or app, capturing every click, scroll, and action for analysis and troubleshooting. These capture the whole user experience making them very valuable for product insight, however there are too many to manually evaluate. So if a human can't do it, why not AI?

How we built it

Loopy is an AI agent that's really good at understanding DOM snapshots of session replays, it only cares about relevant experiences: a user dropping off from onboarding, bugs, and API call fails for example. It will write up descriptions to these issues with the corresponding session replay and display it for you to evaluate on our platform

Challenges we ran into

Although session replays are shown as videos they're actually made up of multiple snapshots of the DOM compiled to together to appear as a video, this data is stored as a large json file. Thus making the processing more complicated than expected

Accomplishments that we're proud of

Loopy is autonomous in distinguishing from relevant and irrelevant session replays, identifying issues and writing corresponding descriptions

What we learned

A lot about session replay data

What's next for Loopy AI

We plan on integrating Loopy in Slack to act as a new member of the product team. He can give weekly summaries of relevant user experiences and ping the team on repeated bugs. Furthermore, we look to make Loopy more autonomous, by also giving him access to Github to create issues for the engineering team.