Evolis

AI Business Rescue – Transforming struggling businesses with smart AI insights.

Comment

Inspiration

Evolis was born from a simple yet powerful observation: every year, thousands of businesses fail not because of fatal flaws, but because they lack access to expert strategic guidance at critical moments. We saw companies struggling with cash flow, declining customer satisfaction, and mounting debt situations that could be reversed with proper analysis and a clear recovery roadmap.

The vision was to democratize business rescue consulting. Instead of expensive consultants being accessible only to large corporations, we wanted to create a platform powered by AI that could:

  • Quickly analyze a company's financial and operational health
  • Identify root causes of business decline
  • Generate actionable recovery strategies tailored to each business
  • Make professional business rescue assistance available to SMEs and startups

Inspired by the real-world pain points of entrepreneurs, i set out to build Evolis as a full-stack SaaS application that bridges the gap between business distress and strategic recovery.

What it does

Evolis is a comprehensive AI-powered business analysis and recovery platform designed to help companies in financial difficulty:

Core Features:

  • Smart Business Analysis: Companies submit detailed information about their operations revenue trends, debt levels, industry challenges, and customer sentiment

  • AI-Powered Diagnosis: Powered by Amazon Nova on AWS, our system analyzes the business holistically and generates:

    • Risk Level Classification (Low, Medium, High)
    • Identified Problems - Deep analysis of root causes
    • Recovery Plan - Step-by-step strategic roadmap
    • Specific Recommendations - Actionable next steps tailored to the industry and context

  • Health Scoring System: A proprietary algorithm computes a health score (0-100) based on:

    • Risk level assessment
    • Revenue trend changes
    • Debt burden ratio
    • Market sentiment analysis

  • Multi-Role Dashboard:

    • Companies can submit analyses, track their health scores, and monitor recovery progress
    • Employees/Consultants can browse and analyze multiple companies
    • Administrators oversee the entire platform, manage user access, and track system health

  • PDF Support: Upload financial documents, reports, or statements for enhanced analysis

  • Real-Time Notifications: High-risk cases trigger immediate notifications to relevant personnel

  • Analysis History & Filtering: Full search, filter, and historical tracking capabilities

How i built it

Evolis was built using modern, production-ready technologies with a focus on security, scalability, and user experience:

Frontend Stack:

  • React + Vite - Fast, component-based UI with hot module reloading
  • Tailwind CSS - Utility-first styling for responsive, polished design
  • React Router - Client-side routing for seamless navigation
  • Firebase Authentication - Secure, OAuth-integrated user management
  • Context API - State management for authentication and user sessions

Backend Stack:

  • Node.js + Express - High-performance REST API with middleware architecture
  • PostgreSQL - Robust relational database for analyzing data integrity
  • AWS Bedrock + Amazon Nova - Advanced LLM for intelligent business analysis
  • Firebase Admin SDK - Server-side authentication and user management
  • Multer - File upload handling for PDF documents
  • PDF-Parse - Document processing and text extraction
  • Nodemailer - Notification system for alerts and communications

Architecture Decisions:

  1. JWT Tokens - Stateless authentication for scalability
  2. Role-Based Access Control (RBAC) - Three distinct user tiers (employee, company, personnel)
  3. Database Migrations - Version-controlled schema evolution
  4. Modular Route Structure - Separated auth, analysis, and support routes
  5. Middleware Pattern - Consistent authentication and authorization checks
  6. Async Processing - Efficient handling of AI analysis requests

Deployment:

  • Vercel configuration for both frontend and backend (serverless architecture)
  • Environment-based configuration for multi-environment support (dev, staging, production)

Challenges i ran into

  1. AI Integration Complexity

    • Challenge: Integrating AWS Bedrock and Amazon Nova required understanding both the SDK and prompt engineering for consistent, high-quality business analysis outputs
    • Solution: Created a dedicated novaService module with fine-tuned prompts that extract structured outputs (risk levels, problems, recovery plans) reliably

  2. Authentication & Authorization Across Roles

    • Challenge: Implementing three distinct user roles (employee, company, personnel) with different data access levels was complex. Companies should only see their own analyses, while personnel can see all analyses
    • Solution: Built middleware with authorizeRoles() that validates JWT tokens and enforces row-level security at the database query level

  3. PDF Processing & Data Extraction

    • Challenge: Many businesses submit PDF documents, but extracting meaningful data from varying PDF formats and layouts was unreliable
    • Solution: Integrated pdf-parse for text extraction and designed the UX to allow users to input structured data alongside optional PDFs

  4. Real-Time Notifications

    • Challenge: Detecting high-risk cases and notifying relevant personnel quickly required building a notification service with low latency
    • Solution: Implemented notificationService using Nodemailer with triggers on high-risk analysis results

  5. Database Schema Evolution

    • Challenge: Starting with one schema and evolving it as features were added (Firebase UID support, additional analysis fields)
    • Solution: Implemented numbered migration scripts (001_init.sql, 002_add_firebase_uid.sql) for reproducible schema changes

  6. Performance with Large Analysis Datasets

    • Challenge: As analyses accumulated, filtering and searching through thousands of records needed to remain fast
    • Solution: Optimized queries with proper indexing and pagination, built dynamic SQL filters based on user role and query parameters

  7. Security & Sensitive Data

    • Challenge: Handling financial company data required strict access controls and secure password storage
    • Solution: Used bcryptjs for password hashing, JWT for stateless auth, and environment variables for sensitive credentials

Accomplishments that i'm proud of

Built a complete, deployable full-stack application - From authentication to AI analysis, Evolis is production-ready with proper error handling and validation

Implemented sophisticated role-based security - Three user tiers with granular data access controls that enforce business logic at every level

Integrated cutting-edge AI - Successfully leveraged Amazon Nova on AWS Bedrock to generate intelligent, structured business analysis outputs

Designed an intuitive user experience - Clean, modern UI with real-time feedback and responsive design that works on desktop and mobile

Created a scalable architecture - Modular codebase with separation of concerns (routes, services, middleware, data layers) that can scale horizontally

Built database migration system - Proper version control of schema changes, allowing team collaboration and safe deployments

Implemented real-world features - PDF uploads, health score calculations, notification system, advanced filtering—features real businesses need

Deployed to production - Configured for Vercel serverless deployment with environment management

What i learned

React & Modern Frontend Development

  • Building responsive UIs with Tailwind CSS
  • Managing state with Context API
  • Protecting routes and handling authentication flows
  • Optimizing Vite builds for fast load times

Node.js Backend Best Practices

  • Structuring Express applications with middleware chains
  • Handling file uploads and processing
  • Building role-based authorization at the API level
  • Error handling and validation strategies

Database Design & Migrations

  • Designing PostgreSQL schemas for complex business logic
  • Implementing proper migrations for safe schema evolution
  • Writing parameterized queries to prevent SQL injection
  • Optimizing queries for performance

AI/LLM Integration

  • Prompt engineering for consistent, structured outputs
  • Working with AWS Bedrock and large language models
  • Handling async AI processing without blocking the user
  • Parsing and validating LLM responses

Security & Authentication

  • JWT token handling and validation
  • Password hashing and secure storage
  • Firebase authentication integration
  • Row-level security enforcement

Full-Stack Development

  • Coordinating frontend and backend development cycles
  • Debugging across the entire stack
  • Testing and validating API contracts
  • Deployment considerations for serverless architectures

What's next for Evolis

Phase 2 - Enhanced AI Capabilities

  • Implement multi-model analysis (comparing Nova with other models)
  • Add industry-specific analysis templates
  • Build longitudinal tracking to monitor recovery progress over time
  • Create predictive models for business outcome forecasting

Phase 3 - Collaboration Features

  • Enable communication between consultants and companies
  • Build commenting and note-sharing on analyses
  • Add document collaboration for recovery plan refinement
  • Implement milestone tracking for recovery plan execution

Phase 4 - Advanced Analytics

  • Industry benchmarking to compare companies against peers
  • Trend analysis dashboards with data visualization
  • Automated report generation for stakeholder communication
  • API for third-party integrations (accounting software, CRM systems)

Phase 5 - Market Expansion

  • Mobile application for on-the-go analysis
  • Support for multiple languages and currencies
  • B2B integrations with business advisory firms
  • Educational content and recovery guides by industry

Phase 6 - Monetization & Growth

  • Subscription tiers (Starter, Professional, Enterprise)
  • Premium features: white-label solutions, API access, dedicated support
  • Partner program with accountants and business consultants
  • International expansion in key markets

Technical Roadmap

  • Migrate to TypeScript for better type safety
  • Implement GraphQL for more flexible API querying
  • Add real-time WebSocket support for live updates
  • Build caching layer with Redis for high-traffic scenarios
  • Implement comprehensive analytics and monitoring with observability tools

Evolis represents mycommitment to making professional business rescue accessible to every company, regardless of size. We're just getting started.

Built With

  • amazon-nova
  • aws-bedrock
  • bcryptjs
  • context-api
  • cors
  • dotenv
  • express.js
  • firebase-admin-sdk
  • firebase-sdk
  • jwt
  • multer
  • node.js
  • nodemailer
  • pdf-parse
  • postgresql
  • react
  • react-router
  • tailwind-css
  • vercel
  • vite
Share this project:

Updates