SUNNY LIU

Outreach Discovery

HackathonFintech

Built for the Mastercard challenge at Microsoft LEAD ASPIRE. Outreach Discovery proposes where a partner bank should meet the customers no CRM can see yet, and how to measure whether that outreach turns into lasting engagement.

Research paper (PDF) ↗
Year
2026
Duration
Finals prototype
Hackathon
Microsoft LEAD ASPIRE
Place
2nd overall, Best in Show
Stack
React · Vite · GitHub Copilot
Outreach Discovery demo showing a bank CRM with zero leads for underserved households
Overview

The challenge premise. Bank CRMs start recording at account opening, so cash-reliant communities with no credit history leave no usable data behind before that moment.

Outreach Discovery is our answer to the question the challenge posed. Where should a partner bank run its next outreach program, and why.

The Problem

The people a bank wants to reach are not in any lead list, because a lead only exists after an account opens. Under this framing, the communities with the least data are exactly the ones outreach is meant for.

Method

We wrote a research paper on which area-level signals are associated with 90-day program retention, then applied the findings in an algorithm that measures trust before a credit score exists.

The demo ranks candidate areas and shows the evidence behind every ranking, so nothing is a black box.

Output

A guided clickable demo that walks from zero leads to a ranked area recommendation, a draft pilot brief for a partner bank, and a learning loop that feeds outcomes into the next decision.

OTHER PROJECTSSee all work →