PAYPROOF

Inspiration

Payment failures can leave merchants with little more than a decline code, while customers and automated shopping agents may not know what to do next. We wanted to make those outcomes easier to understand and help merchants respond with clear, evidence-based next steps.

What it does

PAYPROOF is a payment intelligence platform for merchants and commerce operators. It records checkout outcomes, explains likely causes using transaction evidence, and helps distinguish human checkouts from agent-initiated ones. It also supports eligible recovery flows, creates customer response drafts, and surfaces payment incidents for review.

How we built it

We built PAYPROOF as a Next.js and TypeScript web app. Its checkout uses Braintree Sandbox and Hosted Fields, with a simulated checkout available when sandbox credentials are not configured. Groq provides access to OpenAI GPT-OSS models for incident explanations and response drafts, with template-based fallbacks. Prisma and optional Neon Postgres support persistent data, and AG Grid displays payment records in the admin ledger.

Challenges we ran into

A payment result can show what the processor recorded, but it cannot always explain why a payment failed or what a customer intended. We designed the explanations to point back to available evidence, communicate uncertainty, and avoid unsupported claims. We also had to make the sandbox and simulated payment paths clear so nobody mistakes demo transactions for real payments.

Accomplishments that we're proud of

We brought checkout outcomes, incident analysis, merchant cases, and recovery guidance into one workflow. We’re especially proud that PAYPROOF treats AI output as advisory: merchants can review the evidence and drafts, and consequential actions stay under human control.

What we learned

We learned that useful AI explanations need clear source data and visible limits. We also learned that human and agent checkouts can need different recovery paths, and that payment dashboards are more actionable when they connect metrics to specific transactions and cases.

What's next for PAYPROOF

Next, we want to complete and verify a direct PayPal developer sandbox integration, improve the network analytics experience, and add a hosted demo with a short walkthrough video. We also plan to test the full workflow with realistic sandbox scenarios and make it easier for judges and merchants to explore the demo safely.