We will be undergoing planned maintenance on January 16th, 2026 at 1:00pm UTC. Please make sure to save your work.

AIFinRisk

AIFinRisk is an AI platform that tracks credit, ESG, and market risk across every loan in real time, giving banks and investors a clear, reliable view of true loan health.

Comment

LIVE Demo is Available in this LInk. https://github.com/sureshwizard/aifinrisk/releases/tag/v1.0

Inspiration

The global loan market is one of the largest financial markets in the world, yet most banks still manage loan risk using fragmented spreadsheets, delayed reports, and siloed systems. Defaults, ESG failures, and market shocks are often discovered too late, costing billions and reducing trust in the financial system. We were inspired to build AIFinRisk after seeing how modern AI could transform loan portfolios from static records into living, continuously monitored financial assets.

What it does

AIFinRisk is an AI-powered loan intelligence platform that continuously monitors every loan across 24 independent risk dimensions, including: • Credit risk • Early warning risk • ESG and climate risk • Market and interest rate risk • Sector, geographic, and compliance risk • Recovery and stress-test risk These risk engines are combined into a Master Loan Risk Score that gives banks, investors, and regulators a single, real-time view of loan health. Users can: • See the Top 10 most at-risk loans • Monitor portfolio-level risk trends • Track ESG and climate exposure • Ask questions in natural language using the AI assistant

How we built it

We designed AIFinRisk as a cloud-native, spreadsheet-driven AI platform: • Google Sheets acts as a live relational data fabric for loans, borrowers, and risk engines • Python + Google Colab aggregates 24 independent risk models into a master risk layer • Flask APIs provide real-time access to the data • Gemini AI powers the conversational risk assistant • Tableau delivers interactive portfolio dashboards This architecture makes the system transparent, auditable, and easy for financial institutions to adopt without replacing their existing systems.

Challenges we ran into

The biggest challenge was modeling complex banking risk in a way that is both explainable and scalable. Banks need AI, but they also need transparency, auditability, and regulatory trust. We solved this by separating risk into 24 independent engines, which allows each risk to be understood, tested, and validated independently before being combined into a final score.

Accomplishments that we're proud of

• Built a multi-model AI risk fabric covering 24 dimensions of loan risk • Created a real-time Master Loan Risk Score • Delivered a working AI chatbot that explains loan risk in business language • Designed a system that supports ESG, green finance, and regulatory compliance • Built a commercially viable architecture that can scale from spreadsheets to enterprise systems

What we learned

We learned that the future of financial risk is not a single black-box model — it is a transparent, multi-dimensional AI system that combines data, analytics, and human decision-making. By using spreadsheet-native data and explainable AI, we made advanced risk intelligence accessible to both large banks and smaller financial institutions.

What's next for AIFinRisk

Our roadmap includes: • Live integrations with core banking systems • Advanced stress testing and scenario simulation • Loan trading risk labels for secondary markets • Automated ESG and climate risk reporting • Deployment as a secure SaaS platform for banks and credit funds AIFinRisk is designed to become a digital nervous system for the global loan market.

🧠 AIFinRisk – AI Architecture

AIFinRisk is built as a multi-layer AI risk fabric that transforms raw loan data into real-time financial intelligence.

1️⃣ Data Ingestion Layer (Live Financial Inputs) Sources: • Loan contracts • Payment history • Borrower financials • ESG & climate data • Market & interest-rate data All data is stored in Google Sheets as structured, relational tables: borrowers loans payments financials esg_data market_data This acts as the live financial data lake.

2️⃣ AI Feature Engineering Layer (Risk Signal Generation) Python + Google Colab transforms raw data into meaningful signals: • Debt-to-income • Cashflow stress • Missed payments • Sector exposure • Carbon footprint • Market sensitivity These are written into 24 specialized risk tables, one per risk dimension.

3️⃣ Multi-Model Risk Layer (24 Independent AI Engines) Each risk is modeled separately: • Credit • ESG • Climate • Fraud • Market • Sector • Liquidity • Stress testing • Recovery • And more This ensures explainability and regulatory trust.

4️⃣ Master Risk Engine (AI Aggregation Brain) All 24 risk engines are combined into: master_loan_risk This produces: • Final AI risk score • Risk ranking • Intervention priority This is the single source of truth.

5️⃣ Intelligence & Visualization Layer (Decision UI) • Tableau shows: o Portfolio risk o Trends o Top 10 risky loans o ESG exposure • Gemini AI reads the same data and answers: “Which loans are likely to default?” “Which green loans are at risk?”

6️⃣ Action Layer (What banks do with it) • Trigger credit reviews • Restructure loans • Rebalance portfolios • Enforce ESG compliance • Improve loan trading transparency

AI FIN RISK — Live AI Evidence for Judges

For this submission, all results shown in the screenshots are NOT mock data or static outputs. They are generated in real time by the AI FIN RISK platform, which connects:

• Google Sheets (23 interconnected finance risk tables)

• Python & machine-learning risk engine

• AI (GPT / Gemini) financial analyst

• Live web dashboard

Every screenshot included in this Devpost submission represents a live AI query made against the actual financial risk data model — the same system a real bank or FinTech

🧠 AI FIN RISK — Chatbot Question Library (24+)

🔴 Loan-Level Risk

  1. Which loans are currently high risk?
  2. Why is Loan 102 flagged as high risk?
  3. What is the overall risk score of Loan 115?
  4. Which loans are most likely to default in the next 90 days?
  5. Show me loans with both high credit risk and high fraud risk.
  6. Which borrowers have multiple risky loans? ________________________________________

💳 Credit & Cashflow

  1. Which loans have weak repayment capacity?
  2. Which borrowers show declining cash flow?
  3. What loans have poor credit bureau scores?
  4. Which loans have unstable income risk? ________________________________________

🚨 Fraud & Compliance

  1. Which loans show potential fraud signals?
  2. Which customers are linked to suspicious transactions?
  3. Which loans are failing compliance checks?
  4. Which borrowers have unusually high-risk behavior patterns? ________________________________________

🌍 ESG & Climate Risk

  1. Which loans have high ESG risk exposure?
  2. Which borrowers operate in climate-sensitive regions?
  3. Which sectors have the highest sustainability risk?
  4. Which loans are exposed to greenwashing risk? ________________________________________

📊 Portfolio & Concentration

  1. What is the overall risk level of the loan portfolio?
  2. Which sector has the highest risk concentration?
  3. Which geographic regions are most exposed?
  4. How many loans are in high-risk category? ________________________________________

🧪 Stress Testing & Early Warning

  1. Which loans will fail under a stress scenario?
  2. Which borrowers should be flagged for early intervention?
  3. What is the expected default rate if market conditions worsen? ________________________________________

🤖 Executive-Level AI Questions

  1. What are the top 5 risk drivers in my portfolio?
  2. Which actions should I take to reduce losses?

28. Which loans should be reviewed immediately?

Built With

Share this project:

Updates