Go Beyond

Break the Loop. Go Beyond.

Comment

Break the Loop.

Inspiration

The internet is full of self-improvement advice.

There are thousands of videos, articles, posts, and AI-generated suggestions telling us how to become more confident, disciplined, social, productive, or adventurous. But having more information doesn't necessarily lead to taking action.

We wanted to solve a simpler problem:

What if, instead of giving people more advice, we gave them one thing to actually do?

That idea became Go Beyond.

What We Built

Go Beyond is a personalized self-improvement app that helps people step outside their comfort zone through small, actionable challenges.

Instead of giving users a long list of recommendations, Go Beyond learns their starting point through a short baseline assessment and gives them one clear next move.

For example:

"Start a 2-minute conversation with someone you don't normally talk to."

After completing a challenge, the app tracks the action, rewards progress, and adapts future challenges based on the user's behavior.

The core loop is:

Baseline → Challenge → Action → Completion → Reward → Progress → Next Challenge

The goal isn't to tell someone how to change their life.

It's to help them make one meaningful change today.

Personalization

We built a RAG-powered personalization layer behind the challenge system.

Instead of using AI as a chatbot that generates endless suggestions, the system retrieves relevant challenges based on the user's behavioral profile, previous challenges, completion history, and current progression.

The system then selects a challenge appropriate for the user's current level while avoiding excessive repetition or jumps in difficulty.

This allows the experience to become more personalized as the user actually takes action.

Privacy

We also wanted personalization without unnecessarily exposing identity.

The personalization layer works primarily with a minimized behavioral representation rather than directly exposing information such as the user's email or identity.

Identity tells us who you are. Behavior tells us what you need next.

Monetization

Go Beyond uses RevenueCat to power its Pro subscription.

Users can experience their first five challenges for free before reaching the Pro experience.

Pro unlocks:

  • Unlimited personalized challenges
  • Adaptive difficulty
  • Advanced personalization
  • Push Me
  • Advanced progress insights
  • Long-term growth tracking

RevenueCat handles subscription and entitlement state so premium access isn't hardcoded into the application.

How We Built It

Go Beyond is built as a mobile-first application using:

  • React Native
  • TypeScript
  • Expo
  • Supabase
  • RAG / AI personalization
  • RevenueCat

We focused heavily on making progress visible through the UI. Completing a challenge should feel like an achievement rather than simply checking an item off a list.

Challenges

One of the hardest parts wasn't generating challenges — it was deciding which challenge should come next.

A challenge that is too easy becomes meaningless. A challenge that is too difficult can discourage the user.

We therefore had to think about progression, difficulty, repetition, previous behavior, and personalization together.

Another challenge was integrating monetization without making the product feel like it was blocking the user before they understood its value.

What We Learned

We learned that personalization doesn't necessarily require knowing everything about a user.

A small amount of meaningful behavioral information can be enough to make the next action feel relevant.

We also learned that progress itself can be part of the product. When users can see their actions accumulating into something larger, completing the next challenge becomes more meaningful.

What's Next

We want Go Beyond to become increasingly adaptive as users build a history of real-world actions.

The long-term vision is simple:

Less scrolling. Less overthinking. Less noise. More doing.

You don't need more advice.

You need your next move.

Built With

Share this project:

Updates

Submission history