Open-source Revenue Leak Prediction Tool for Web and Mobile Apps

**UT Austin Student Develops AI-Powered Tool to Predict App Issues**

A sophomore at the University of Texas at Austin, Rashid, has created an innovative open-source tool called Rejourney aimed at predicting issues in apps and websites before they occur. Rejourney uses real user session recordings to identify potential problems, allowing developers to fix them before they affect a large number of users. The tool is based on machine learning algorithms and can process thousands of user recordings daily.

Rejourney was initially developed to address issues with Rashid's own campus freebie finder app, which grew rapidly but suffered from onboarding and UX confusion problems. After losing around 340 users due to these issues, Rashid created Rejourney to prevent similar problems from occurring in the future. The tool works by recording user sessions, relating them to critical conversion events, and analyzing the sequence of user interactions to identify potential issues. If a trend is found, Rejourney's Large Language Model (LLM) processes the user recordings to determine the likelihood of a negative outcome and outputs a report with suggested fixes.

Rejourney has already shown promising results, with one user reporting a 30% increase in onboarding after fixing non-stop issues found by the tool. The company has also broken even on costs with its first three paid users, allowing for further development and expansion. With a focus on cost-effectiveness and user privacy, Rejourney aims to become a valuable resource for developers looking to improve their apps and websites.