Landing Page
Landing Page Pitch Elements
Auth0 Authentication Role Based
Product Customer Happiness Index Dashboard
Product Insight View
Customer Loyalty Automation
Technical Ticket Automation
Mass Feedback Call Automation
Eleven Labs Agent Sample Call

A.U.R.A.: The AI-Powered PM Agent for T-Mobile's "Happiness Index"

Inspiration

We didn't start this hackathon with a piece of tech; we started by investigating a user's frustration. We spoke to a family friend, a Product Manager at a major tech company, and asked them, "What's the most chaotic, time-wasting part of your job?" The answer was clear: drowning in endless data (feedback tickets, analytics, user complaints) but having no time to connect it to clear, confident actions.

We also spoke to a parent who works with a major telecom client. They highlighted that in the telecom industry, with its long-term user cycles, the most important drivers for success are consistent feedback and rewarding loyalty.

This sparked our "Aha!" moment. PMs are drowning in data, while the business needs action on feedback and loyalty. What if we built a tool that automates this entire workflow? A tool that could turn customer emotion into actionable insight and give PMs their time back?

What it does

A.U.R.A. (Automated Unified Response Agent) is our answer. It's an AI-powered productivity solution for Product Managers built to solve two challenges at once.

First, A.U.R.A. builds the T-Mobile Customer Happiness Index (CHI)—a real-time, data-driven solution that captures how customers feel. Our "CHI Engine" streams live sentiment, feedback, and network data from four distinct sources:

Internal Customer Feedback
A Real Disaster Weather API (visualcrossing.com)
A T-Mobile Network Outage API
A Social Media Sentiment API

It doesn't just give a score. Our engine calculates a weighted ImpactVector to identify the Primary Cause of any happiness drop (e.g., "Network Outage is 60% of the problem"). This is how we detect issues before they spread.

Second, A.U.R.A. is the PM's "Auto-Pilot." When a PM logs in, they don't just see a dashboard; they see a co-pilot. A.U.R.A. provides an intelligent agent with three contextual, one-click actions to empower teams and accelerate speed to market:

Service Outage Detected (CHI Low): The PM clicks "Raise Tech Support Ticket." Instantly, Gemini Agent #1 acts as an Ops Manager, analyzing the outage data and customer reports to draft a professional, urgent JIRA ticket. This is automated Requirements & Development.
Poor Feedback Detected (CHI Low): The PM clicks "Send Feedback Calls" for deep Customer & Market Research. Our ElevenLabs agent proactively calls a customer (our teammate) with a dynamic script. The new, verbal feedback is transcribed and sent to Gemini Agent #2, which generates a "Product Strategy Memo" complete with root cause, recommendations, and draft user stories.
High Satisfaction Detected (CHI High): To highlight moments of delight, the PM clicks "Generate Loyalty Promotions." Gemini Agent #3 acts as a Marketing Manager, drafting an exciting email to reward loyal customers and drive more business. This is high-speed Go-to-Market Execution.

How we built it

We chose a modern, scalable, full-stack architecture to build this 12-hour demo.

Frontend: A dynamic, responsive dashboard built with React (using Vite) and styled with Tailwind CSS. Hosted on Vercel.
Backend: A high-performance API server built with Python and FastAPI, serving as the "brain" for the CHI Engine and all AI agents. Hosted on Railway.
Database: A PostgreSQL database hosted on Supabase to manage products, user data, and feedback.
AI & APIs:
- Google Gemini API: The core intelligence, used for four distinct agentic tasks (Summarization, Technical Writing, Strategy, Marketing).
- ElevenLabs API: Used for our proactive, real-time feedback call agent.
- Auth0: For secure, enterprise-ready PM login.
- Visual Crossing Weather API: Our real-world data source for disaster alerts.

A Note for the Judges: Our Alignment with Your Tracks

We were extremely intentional in designing A.U.R.A. as a direct, comprehensive solution to your challenges.

For the PNC Bank Challenge

We built A.U.R.A. as the definitive solution to the PNC challenge, reimagining how artificial intelligence can be embedded across the product lifecycle:

Product Strategy & Ideation: Our ElevenLabs-to-Gemini workflow turns a live customer call into a complete strategy memo with recommendations, aligning customer needs with business goals.
Requirements & Development: Our Gemini agents drafted user stories and high-priority technical JIRA tickets, providing AI-assisted clarity.
Customer & Market Research: The core of our CHI Engine is to synthesize customer feedback, competitor activity (social API), and industry trends (disaster API) into a single, actionable insight.
Go-to-Market Execution: The "Generate Loyalty Promotions" agent is a one-click tool to assist in GTM strategy and stakeholder communication.
Automation & Intelligent Agents: We deployed five distinct intelligent agents (four Gemini, one ElevenLabs) to automate repetitive Product Manager workflows and accelerate speed to market.

For the T-Mobile Challenge

We built the "T-Mobile Customer Happiness Index" from the ground up as a real-time, data-driven solution.

Our CHI Engine streams live sentiment, feedback, and network data from four different sources to create one unified, actionable score.
The ImpactVector is our core feature to detect issues before they spread by identifying the Primary Cause of a problem.
Our 3-button "Auto-Pilot" is the literal tool to empower teams and turn customer emotion into actionable insight—transforming a low score into a JIRA ticket or a strategy memo in a single click.

For MLH (Gemini, ElevenLabs, Auth0 & GoDaddy Registry)

Gemini API: We demonstrate the incredible range of Gemini by using it for four completely different, high-value tasks: initial summarization, technical writing (JIRA), creative strategy (Memo), and marketing copywriting (Email).
ElevenLabs API: We created a seamless AI-to-AI workflow, where the output of an ElevenLabs call (a transcript) becomes the input for a Gemini-generated strategy memo, building a rich, immersive experience.
Auth0: We used Auth0 to make our application secure and enterprise-ready. We implemented role-based authentication to ensure Product Managers can only see the product suites they are authorized to manage, securing our application straight out of the box.
GoDaddy Registry: We are hosted on a professional, memorable domain name from GoDaddy Registry, [YOUR_DOMAIN_HERE.AI], making our solution polished and market-ready.

Challenges we ran into

Our biggest challenge wasn't technical; it was strategic. First, we had to truly understand the primary domain concerns of a Product Manager to avoid building a tool they wouldn't use.

Second, we had to connect those concerns to a meaningful "Happiness Index"—not just a vanity metric. Designing the ImpactVector logic, which weights different data sources (like a network outage vs. a social media complaint), was crucial.

Finally, ensuring our AI agents generated correct, actionable insights required significant prompt engineering. A generic prompt gives a generic result. We had to build four distinct, expert personas for our Gemini agents to ensure their outputs were professional, relevant, and immediately usable.

Accomplishments that we're proud of

We are incredibly proud of building a full, end-to-end, agentic workflow that solves a real-world problem. Our system doesn't just find problems; it empowers the PM to fix them in one click.

We're also proud that A.U.R.A. not only makes the product better (by automating issue detection) but also directly drives T-Mobile's business goals. By rewarding loyalty and highlighting moments of delight, our promotion-generation agent gives PMs the ability to directly impact revenue-focused KPIs, not just development velocity.

What we learned

Our biggest takeaway is the power of contextual, agentic AI. We learned that a "one-size-fits-all" AI chatbot is a gimmick. The real power—the future of AI-powered productivity—comes from deploying specialized, intelligent agents (like our four Gemini personas) that are given specific context (from the CHI engine) to perform a specific, high-value task.

What's next for A.U.R.A.

This 24-hour project is just the beginning. The vision is to make the loyalty and product promotions more strategic by integrating with real marketing and CRM platforms. We also plan to expand A.U.R.A. into a global Product Management automation platform, integrating directly with T-Mobile's real internal network APIs and JIRA to create a truly seamless "Auto-Pilot" for the entire product lifecycle.