Skip to main content
Back to The AI Commerce Gallery - Hackathon | Walt Disney Family Museum Gallery

Project spotlight

FinalistRank #4

PACT

by Team PACT

Agents negotiate. You decide.

PACT

Demo Video

About This Project

Imagine a store is closed, but a shopper’s agent is still asking for a deal. PACT gives the merchant an agent that can answer.

PACT enables shopper and merchant agents to negotiate price, delivery, and returns within their owners’ boundaries. The shopper sets a budget and approval threshold. The merchant protects its pricing rules, minimum profit, inventory, and return policies while keeping private costs private.

Our headphone demo compares offers from two merchants. The agents reach an agreement, then an authority gate checks whether they have permission to proceed. A purchase can fit the shopper’s budget and still require human approval. PACT pauses at that boundary.

BAND provides the negotiation rooms, ZooWork provides the merchant runtime, Jev checks authority, Tavily verifies competitor prices, and Moss preserves preferences and outcomes. The same conversation continues through delivery and return negotiations.

We built PACT to help merchants convert shoppers while protecting their economics. The demo clearly labels live integrations and local fallbacks. Checkout, delivery events, and return execution are simulated.

Repository

HTML98%Python1.9%CSS0.1%JavaScript0%
Last commit 19 days ago