GooseOnCraze Agent

An agent that shows response with interactive UI and connects to any MCP server to automate your CRM tasks, trigger complex workflows, and even initiate agent calls.

Inspiration

Traditional enterprise resource planning (ERP) systems have long been plagued by clunky interfaces, steep learning curves, and disconnected workflows. We were inspired to create a solution that reimagines how businesses interact with their data and processes through natural language. By combining the intuitive nature of conversational AI with the versatility of multi-chain prompting (MCP), we set out to build an ERP system that adapts to users rather than forcing users to adapt to it.

What it does

NewAge ERP is a next-generation business management platform powered by AI that enables users to:

Execute complex business tasks through simple natural language requests
View and interact with data through dynamically generated UI components (tables, forms, charts)
Connect to multiple external services via MCP servers (databases, email services, web scrapers, workflow engines)
Automate end-to-end business processes combining multiple tools (e.g., scrape website → send targeted email → initiate follow-up call)
Generate real-time analytics and visualizations from enterprise data

The system eliminates the need to navigate complex menus or learn specialized query languages. Users simply explain what they need, and the AI handles the rest, presenting results in the most appropriate interactive format.

How we built it

We built NewAge ERP as a modern web application with a focus on modularity and extensibility:

Frontend: React with Next.js, using a component-based architecture with shadcn/ui for consistent design
UI Generation: Dynamic UI component system that renders appropriate interfaces based on AI responses
Backend: Node.js with Express, structured to handle various data sources and external services
AI Integration: Large language model integration for natural language understanding and response generation
MCP Framework: Custom Multi-Chain Prompting system to connect with external tools and services
Mock Services: Comprehensive mock data and service simulations for development and demonstration

The architecture is centered around a conversational interface that translates user requests into structured operations, executes them across the appropriate services, and returns results in interactive UI components.

Challenges we ran into

Building NewAge ERP presented several significant challenges:

Dynamic UI Generation: Creating a system that could reliably generate appropriate UI components based on varied AI responses
Type Safety vs Flexibility: Balancing TypeScript's type safety with the need for dynamic, runtime-determined UI structures
MCP Integration: Designing a secure, extensible system for connecting to external services with proper authentication
State Management: Handling complex application state across asynchronous operations and multiple connected services
User Experience: Ensuring the conversation-to-UI flow felt intuitive and predictable despite the complexity beneath

We solved these through careful architecture decisions, extensive type definitions, and continuous user experience testing.

Accomplishments that we're proud of

Our team is particularly proud of several accomplishments with NewAge ERP:

Creating a seamless integration between natural language and interactive UI components
Developing a flexible MCP architecture that can connect to virtually any external service
Building an end-to-end workflow system that chains multiple operations across different services
Implementing a visually appealing, responsive design that works across devices
Maintaining strong type safety throughout the codebase despite the dynamic nature of the application

What we learned

This project provided valuable learning opportunities:

Advanced techniques for dynamic UI generation based on data types and context
Strategies for integrating LLMs with traditional business applications
Best practices for type safety in highly dynamic TypeScript applications
Approaches to designing multi-step workflows with proper error handling and recovery
Methods for balancing immediate user feedback with background processing

What's next for NewAge ERP

The future roadmap for NewAge ERP includes:

Advanced Customization: Allowing users to create and save custom workflows and UI preferences
Enhanced Data Visualization: More sophisticated chart types and interactive dashboards
Expanded MCP Ecosystem: Pre-built connectors for popular business services and platforms
Mobile Application: Native mobile apps for iOS and Android
Enterprise Integrations: Connectors for popular enterprise systems like Salesforce, SAP, and Microsoft Dynamics
Knowledge Graph: Building a domain-specific knowledge graph from enterprise data to improve AI responses
Collaborative Features: Multi-user workspaces and collaboration tools

We're excited to continue developing NewAge ERP into a comprehensive platform that revolutionizes how businesses interact with their data and processes.

Features

Intelligent Agent-based Interface: Interact with your ERP system through natural language
Multi-modal UI Components: Agent responds with appropriate UI for different tasks
Real-time Updates: WebSocket integration for live agent progress updates
CRM Module: Manage todos and leads
Expense Management: Track and approve expenses

Architecture

Frontend (Next.js)

Next.js 15: Modern React framework
Shadcn/UI: Beautiful component library
TypeScript: Type-safe code
WebSocket: Real-time updates from agent

Backend (FastAPI)

FastAPI: Modern Python web framework
LangGraph: Multi-step agent orchestration
SQLite: Local database
Pydantic: Type validation
OpenAI Integration: Intent parsing with GPT-4o-mini (API key pre-configured)