Inspiration

AI agents are becoming capable of making business decisions, but giving autonomous systems unrestricted access to company funds creates serious risks. We wanted to solve one critical problem: How can businesses trust AI to negotiate and make purchases without losing financial control?

What it does

PACTPAY is an AI-powered B2B procurement and payment platform. It enables AI agents to negotiate with suppliers, compare market prices using Channel3, and identify better purchasing deals.

Before any payment, Microsoft's Z3 solver mathematically verifies budgets, pricing, and business policies. Cryptographic authorization ensures that only approved transactions reach PayPal for execution.

Our principle: Let AI negotiate. Let mathematics verify. Let businesses control their money.

How we built it

We built PACTPAY using Python, FastAPI, Next.js, Microsoft Z3, Channel3 APIs, and PayPal Orders API.

The system combines autonomous negotiation, market-price benchmarking, formal constraint verification, cryptographic clearance certificates, and controlled payment execution. We also developed a procurement dashboard, red-team testing console, and audit mechanisms to improve transparency and security.

Challenges we ran into

Integrating multiple APIs and handling inconsistent responses.
Designing reliable negotiation logic while preventing budget violations.
Translating business policies into mathematically verifiable constraints.
Securing the boundary between AI decision-making and actual payment execution.
Handling PayPal Sandbox configuration, currency compatibility, and transaction workflows.

Accomplishments that we're proud of

Built an end-to-end AI-driven procurement and payment prototype.
Integrated Z3-based formal verification into the purchasing workflow.
Implemented cryptographic authorization for payment clearance.
Connected supplier price intelligence through Channel3.
Developed security-focused scenarios to test budget violations, manipulated pricing, and unsafe agent behavior.
Demonstrated how autonomous AI negotiation can work alongside deterministic financial controls.

What we learned

We learned that building trustworthy AI is not just about making agents smarter. It is about designing systems that constrain their authority, verify their decisions, and maintain accountability. Formal methods and cryptography can turn AI-generated decisions into controlled, auditable business actions.

What's next for PACTPAY

We aim to evolve PACTPAY into a broader autonomous procurement infrastructure for businesses, with multi-supplier marketplaces, enterprise policy management, stronger supplier intelligence, expanded payment integrations, and production-grade auditability.

Our long-term vision is to make autonomous business transactions secure, verifiable, and accountable by design.