GuardPay

Secure and Trustworthy Agentic Commerce

Inspiration

AI agents are becoming increasingly capable of shopping and making decisions on behalf of users. However, giving AI agents access to money introduces significant trust and security challenges.

We wanted to build a system that enables users to experience autonomous shopping without losing control over their spending. This inspired us to create GuardPay, a secure, policy-driven payment layer designed specifically for agentic commerce.

What It Does

GuardPay allows users to shop using natural language. Its AI-powered agent discovers and ranks products, builds shopping carts, and negotiates prices with seller agents.

Before any purchase is made, a deterministic policy engine evaluates spending limits and approval requirements. Payments are processed through PayPal Sandbox, while audit logs and real-time event tracking ensure transparency and accountability.

How We Built It

We developed the frontend using Next.js, TypeScript, Tailwind CSS, and shadcn/ui.

Google Gemini Flash and Groq, integrated through the Vercel AI SDK, power the AI workflows. Channel3 API enables product discovery, while Neon PostgreSQL and Drizzle ORM handle data storage and management.

PayPal Orders API (Sandbox) powers payment processing. Additionally, we implemented a deterministic policy engine, human approval workflows, and Server-Sent Events (SSE) for real-time updates.

Challenges We Ran Into

One of our biggest challenges was balancing AI autonomy with payment security. Since AI models can produce unpredictable decisions, we needed to ensure that financial authorization remained independent of AI-generated decisions.

We also faced challenges in designing reliable buyer-seller negotiations, maintaining consistent cart and payment states, and coordinating multiple external API integrations.

Accomplishments That We're Proud Of

We are proud to have built an end-to-end agentic commerce platform where AI agents can discover products, negotiate deals, and prepare purchases while users remain in control of their money.

Our biggest achievement is successfully combining autonomous AI workflows with deterministic spending policies, human approval mechanisms, and transparent audit trails.

What We Learned

We learned that building trustworthy AI agents requires more than intelligent models. Critical operations, especially payment authorization, must rely on deterministic rules rather than AI decisions alone.

We also gained valuable hands-on experience with multi-agent workflows, third-party API integrations, payment processing systems, real-time communication, and full-stack development.

What's Next for GuardPay

We plan to expand GuardPay by introducing additional payment providers, stronger fraud detection capabilities, customizable security policies, improved seller-agent integrations, and production-ready payment infrastructure.

Our long-term vision is to make autonomous commerce secure, transparent, and trustworthy for everyone.