Inspiration

Business procurement is often a fragmented process. Businesses need to coordinate with multiple suppliers, compare quotations, evaluate purchasing options, obtain approvals, and manage payments across different platforms.

While generative AI can help businesses make smarter purchasing decisions, connecting those decisions to actual financial transactions remains a challenge.

I wanted to explore a more practical application of AI — one that goes beyond generating recommendations and becomes part of a real business workflow.

This inspired me to build Procura, an AI-powered B2B procurement copilot that connects intelligent purchasing assistance with PayPal-powered financial settlement.

My vision was to create a unified experience where AI helps businesses understand what to buy, compare where to buy it, and proceed towards payment through an integrated payment infrastructure.

What it does

Procura transforms natural-language procurement requirements into AI-assisted purchasing decisions and connects approved procurement selections to PayPal payment processing.

Its key capabilities include:

  • AI-Powered Procurement: Uses Google Gemini to interpret natural-language purchasing requirements, analyze procurement information, and assist with intelligent purchasing decisions.
  • Multi-Supplier Evaluation: Organizes supplier quotations, pricing, availability, and product information to help businesses evaluate multiple vendors.
  • AI-Assisted Recommendations: Helps identify suitable supplier options based on purchasing requirements and available information.
  • Interactive AG Grid: Provides an interactive interface for viewing, organizing, and comparing supplier quotations and procurement data.
  • PayPal Payment Integration: Uses PayPal Orders API v2 to create orders and facilitate payment processing through the PayPal Sandbox environment.
  • Human-in-the-Loop Approval: Allows users to review AI-assisted recommendations and approve purchasing decisions before initiating payments.
  • Transaction Verification: Connects payment processing with transaction verification and procurement records.
  • Unified Dashboard: Brings procurement operations, AI assistance, supplier evaluation, and payment workflows into one interface.

What makes Procura different is how these capabilities work together.

Instead of stopping at an AI-generated recommendation, Procura connects procurement intelligence to a practical financial settlement workflow.

How I built it

I independently designed and developed Procura by combining generative AI, interactive data management, and payment infrastructure.

Technology Stack:

  • React + Vite: Frontend application and interactive user interface.
  • Google Gemini API (@google/genai): Powers natural-language understanding, procurement analysis, and AI-assisted purchasing workflows.
  • AG Grid Community: Handles interactive supplier comparisons and structured procurement data presentation.
  • PayPal Orders API v2: Supports payment order creation and payment processing.
  • PayPal Sandbox: Provides a simulated environment for testing the payment lifecycle without transferring real money.
  • JavaScript, HTML, CSS: Application logic, structure, and styling.

The implementation brings together three major layers:

1. Intelligence Layer – Google Gemini

Gemini interprets purchasing requirements, processes procurement information, and assists with supplier evaluation and recommendations.

2. Procurement Data Layer – AG Grid Community

AG Grid provides an interactive interface for presenting supplier quotations, prices, quantities, and comparison information, helping users evaluate their purchasing options.

3. Financial Execution Layer – PayPal

Once a user reviews and approves a procurement decision, Procura uses PayPal Orders API v2 to create the corresponding payment order and initiate the checkout workflow.

The user authorizes the payment through the PayPal Sandbox checkout experience. The backend then handles payment capture and transaction verification.

I also implemented duplicate-capture protection to help prevent repeated payment processing.

This architecture connects AI-assisted purchasing decisions with a structured payment workflow while keeping the final purchasing authorization in the user's hands.

Challenges I ran into

One of my biggest challenges was connecting AI-generated procurement insights with a reliable payment workflow.

Some of the key challenges included:

  • AI Integration: Making Gemini's natural-language understanding and AI-assisted recommendations work within a structured procurement process.
  • Supplier Data Management: Presenting information from multiple suppliers in a clear and interactive format using AG Grid.
  • PayPal Orders API Integration: Understanding the payment lifecycle, implementing order creation, handling checkout authorization, and processing payment capture through the Sandbox environment.
  • Human Approval and Payment Coordination: Ensuring that AI-assisted recommendations, user approval, and payment execution remained connected without allowing AI to independently authorize transactions.
  • Transaction Reliability: Handling payment responses, verifying transaction details, and preventing duplicate capture attempts.
  • Integration and Debugging: Resolving package compatibility issues, frontend-backend coordination, and API integration challenges.

These challenges taught me that building an AI-powered business application requires more than integrating an AI model. Every component must work together reliably to deliver a meaningful end-to-end experience.

Accomplishments that I'm proud of

I'm proud to have independently transformed an idea into a functional procurement-to-payment prototype within the hackathon timeline.

My key accomplishments include:

  • Successfully integrating Google Gemini into a practical B2B procurement application.
  • Building AI-assisted procurement workflows that go beyond simple question-and-answer interactions.
  • Developing an interactive multi-supplier comparison experience using AG Grid Community.
  • Integrating PayPal Orders API v2 and demonstrating payment processing through the PayPal Sandbox.
  • Connecting procurement selection, human approval, payment execution, and transaction verification into one workflow.
  • Developing a unified application that brings AI intelligence and financial infrastructure together.

Most importantly, I'm proud that Procura demonstrates a practical use of generative AI where intelligence leads to meaningful business actions rather than stopping at recommendations.

What I learned

Building Procura gave me valuable hands-on experience in developing and integrating an AI-powered business application.

I learned how to:

  • Integrate Google Gemini into real-world application workflows.
  • Work with AI-assisted procurement analysis and structured supplier information.
  • Build interactive business data interfaces using AG Grid.
  • Understand and implement the PayPal Orders API v2 payment lifecycle.
  • Use PayPal Sandbox to test simulated payment transactions.
  • Coordinate AI responses, frontend interactions, backend operations, and external payment APIs.
  • Debug complex integration issues and improve application reliability.

My biggest takeaway was that the real value of AI comes from connecting intelligence with tools, structured data, and actions that solve actual problems.

What's next for Procura

I see Procura as a foundation for a more intelligent and automated procurement ecosystem.

My future plans include:

  • Advanced Supplier Intelligence: Improving supplier evaluation through richer comparison criteria and more contextual recommendations.
  • Automated Procurement Workflows: Reducing manual intervention between purchasing requirements, supplier evaluation, and approvals.
  • Smarter AI Agents: Introducing more advanced agentic capabilities for procurement research, supplier analysis, and workflow coordination.
  • Expanded Payment Automation: Exploring more sophisticated financial workflows and multi-supplier settlement capabilities.
  • Procurement Analytics: Improving visibility into purchasing patterns, spending, supplier performance, and cost savings.
  • Production Readiness: Strengthening security, transaction reliability, scalability, and integration capabilities for real-world business adoption.

My long-term vision is to transform procurement from a fragmented manual process into an intelligent, connected journey.

Procura – Where AI-Powered Decisions Meet Seamless Payments.

Built With

Share this project:

Updates

Submission history