RootCause AI

RootCause AI transforms STEM education through AI-assisted investigations, enabling students to analyze evidence, identify root causes, and think like engineers through real-world failure scenarios.

Comment

Inspiration

What happens when a drone suddenly crashes mid-flight?

Or when a robotic arm stops working in the middle of an assembly operation?

Or when a chemical reactor experiences a dangerous pressure spike?

In the real world, engineers don't immediately know the answer. They investigate. They collect evidence. They form hypotheses. They connect cause and effect. They identify root causes.

Yet in most classrooms, students are rarely taught how to think through failures. They are taught formulas, definitions, and solutions—but not the investigative reasoning process that engineers and scientists use every day.

We wanted to change that.

RootCause AI was inspired by a simple question:

What if students could learn STEM by solving mysteries instead of memorizing answers?

Our goal was to create a platform that transforms learners into investigators, allowing them to think like engineers, analyze evidence like scientists, and develop critical reasoning skills through realistic failure scenarios.

What it does

RootCause AI is an AI-powered STEM investigation platform that turns technical failures into interactive learning experiences.

Users are presented with realistic engineering and scientific incidents such as:

  • Drone crashes
  • Robot malfunctions
  • Chemical reactor failures
  • Communication system breakdowns

Instead of being given the answer, students must:

  1. Examine evidence
  2. Identify patterns
  3. Build a hypothesis
  4. Explain their reasoning
  5. Submit their investigation

The AI then evaluates their reasoning process, identifies strengths and weaknesses, predicts the most likely root cause, generates confidence scores, and visualizes the failure chain that led to the incident.

Rather than rewarding memorization, RootCause AI rewards structured thinking.

How we built it

We built RootCause AI as a modern full-stack web application focused on educational AI and interactive investigation workflows.

Frontend

  • Next.js 16
  • React 19
  • TypeScript
  • Tailwind CSS

AI & Evaluation Layer

  • Google Gemini API
  • Custom prompt engineering
  • Reasoning evaluation engine
  • Root cause detection logic

Core Systems

  • Evidence management system
  • Investigation workflow engine
  • AI feedback generator
  • Timeline reconstruction engine
  • Root cause chain visualizer
  • Scoring and confidence framework

The platform combines AI reasoning with structured engineering logic to create an educational experience that is both engaging and technically meaningful.

Challenges we ran into

The hardest challenge was not building the interface—it was teaching AI to evaluate reasoning.

Unlike traditional educational platforms where answers are either right or wrong, investigation-based learning exists in a gray area. A student can reach the correct conclusion using poor reasoning, or reach an incorrect conclusion using strong analytical thinking.

We had to design a system capable of evaluating:

  • Evidence usage
  • Logical connections
  • Cause-and-effect reasoning
  • Engineering thought process
  • Investigation completeness

We also faced challenges in:

  • Designing realistic STEM cases
  • Building meaningful root cause chains
  • Creating educational AI feedback
  • Balancing flexibility with consistent scoring
  • Optimizing production deployment

Accomplishments that we're proud of

We're proud that RootCause AI became much more than a simple AI chatbot.

We successfully built:

✅ A complete STEM investigation platform

✅ AI-powered reasoning evaluation

✅ Root cause analysis workflows

✅ Investigation timeline reconstruction

✅ Failure chain visualization

✅ Multi-domain STEM case library

✅ Modern and responsive user experience

Most importantly, we created a system that encourages students to think critically rather than simply search for answers.

What we learned

Building RootCause AI taught us that educational AI becomes significantly more powerful when it guides thinking instead of replacing it.

Throughout the project we learned:

  • Prompt engineering for educational applications
  • Designing AI-assisted learning systems
  • Building scalable Next.js applications
  • Structuring reasoning evaluation frameworks
  • Creating engaging STEM experiences
  • Turning complex engineering concepts into interactive learning

We also discovered how powerful investigation-based learning can be when combined with AI.

What's next for RootCause AI

This project is only the beginning.

Our vision is to evolve RootCause AI into a complete STEM investigation ecosystem.

Future plans include:

  • Student accounts and progress tracking
  • AI-generated investigation scenarios
  • Team-based investigations
  • Educator dashboards
  • Achievement systems and leaderboards
  • Personalized AI tutoring
  • Advanced analytics
  • Industry-inspired case studies
  • Downloadable investigation reports
  • Classroom integration

We believe the future of STEM education is not just teaching students what to think—it is teaching them how to think.

RootCause AI is our step toward that future.

Built With

Share this project:

Updates