InsightForge

Upload your data. Discover trends, predict outcomes, and make better business decisions with AI.

Comment

Inspiration

Small and medium-sized businesses generate large amounts of data through sales transactions, inventory operations, and customer interactions. However, many SMBs still rely on spreadsheets and manual reporting because advanced business intelligence tools are often expensive, complex, or require dedicated analytics teams.

We wanted to build a platform that could bridge this gap by allowing businesses to upload their existing data and instantly receive meaningful insights, forecasts, and recommendations without needing technical expertise.

This idea became InsightForge.

What it does

InsightForge is an AI-powered Business Intelligence and Decision Support Platform designed specifically for small and medium-sized businesses.

Users can upload sales, inventory, and customer datasets directly into the platform. InsightForge automatically detects dataset types, maps important business metrics, and generates actionable insights.

Key capabilities include:

  • Automatic dataset detection and column mapping
  • Business Health Score calculation
  • Sales, inventory, and customer analytics
  • Forecast readiness assessment
  • Risk identification and anomaly detection
  • AI-generated executive summaries and recommendations
  • Historical business performance tracking

The goal is simple: help SMBs make data-driven decisions without needing a dedicated analytics team.

How we built it

InsightForge was built using a modern full-stack architecture.

Frontend

  • Next.js
  • React
  • Tailwind CSS
  • Vercel Deployment

Backend

  • Next.js API Routes
  • Prisma ORM

Database

  • Amazon Aurora PostgreSQL

AI Layer

  • Google Gemini API

One of the most important architectural decisions was selecting Aurora PostgreSQL as the database layer.

InsightForge manages highly related business entities including users, uploaded datasets, business health scores, forecasts, and AI-generated reports. These relationships require transactional consistency, structured querying, and historical tracking.

Aurora PostgreSQL allowed us to build a relational foundation that supports long-term business intelligence rather than one-time CSV analysis.

We also built a custom dataset detection and column mapping engine that automatically identifies important fields such as revenue, product information, inventory metrics, customer identifiers, and forecasting signals even when businesses use different column naming conventions.

Challenges we ran into

One of the biggest challenges was handling real-world business data.

Different businesses structure their datasets differently. Revenue may appear as Revenue, Sales, Amount, Total Value, or other custom fields. Inventory and customer datasets also vary significantly across industries.

To solve this problem, we built an intelligent dataset detection and column mapping system that evaluates uploaded datasets and automatically identifies supported business metrics.

Another challenge was designing a flexible analytics pipeline that could work even when businesses only upload partial datasets. InsightForge can still generate meaningful insights from sales-only data while providing richer analytics when inventory and customer datasets are also available.

We also spent significant effort ensuring that AI-generated recommendations were grounded in actual business metrics rather than producing generic responses.

Accomplishments that we're proud of

  • Built a complete end-to-end business intelligence platform from scratch
  • Successfully integrated Amazon Aurora PostgreSQL with a Vercel-hosted application
  • Developed automatic dataset detection and intelligent column mapping
  • Created a Business Health Score system that evaluates multiple dimensions of business performance
  • Implemented forecasting, anomaly detection, and risk analysis features
  • Built AI-generated executive reports that convert raw data into actionable recommendations
  • Designed a platform that can provide value even when businesses have incomplete datasets

What we learned

Building InsightForge reinforced how important data modeling is for business intelligence systems.

We learned that:

  • Database design directly impacts analytical capabilities
  • Real-world business datasets are rarely standardized
  • AI recommendations become significantly more useful when combined with structured business data
  • Business users care more about actionable decisions than raw analytics
  • Relational databases remain critical for systems that manage interconnected business entities

We also gained valuable experience combining Vercel's deployment platform with AWS database infrastructure to build scalable data-driven applications.

What's next for InsightForge

We plan to continue evolving InsightForge into a comprehensive decision intelligence platform.

Future improvements include:

  • Real-time integrations with Shopify, WooCommerce, and ERP systems
  • Automated inventory optimization recommendations
  • Advanced customer segmentation and churn prediction
  • Real-time anomaly detection
  • Multi-user collaboration and role-based permissions
  • Industry-specific analytics modules
  • Enhanced forecasting models powered by historical business performance

Our vision is to help every SMB access the same level of business intelligence that is traditionally available only to large enterprises.

Share this project:

Updates

Submission history