The Credit Analysis Report Tool is designed to streamline the evaluation of a borrower’s financial health by automating data extraction, ratio computation, and risk assessment. It converts raw financial statements into clear, decision-ready analytics, enabling faster and more consistent credit profiling. This use case demonstrates how AI-driven automation enhances accuracy, reduces manual effort, and strengthens credit decision-making for financial institutions.X

Problem statement

When an organisation seeks a credit or working capital facility from a bank, the lending process requires an independent assessment of its financial strength and repayment capacity. Banks rely heavily on external credit ratings before approving or enhancing loan limits. To obtain this rating, the organisation must coordinate with a credit rating agency, provide extensive financial and business information, and undergo a detailed evaluation. This process is often time-consuming, data-intensive, and involves multiple back-and-forth interactions, creating delays and inconsistencies in credit assessment.

EXISTING WORKFLOW

Existing Workflow – Key Points

1. Financial Statements received in PDF format
2. Annual financial statements (Profit & Loss, Balance Sheet, Cash Flow) are typically shared in PDF, scanned, or unstructured formats.
3. Manual copy-paste into working templates
4. Teams manually extract data line-by-line from PDFs and paste them into Form-II (Operating Statement) and Form-III (Balance Sheet Analysis).
5. This process is repetitive, time-consuming, and prone to human error.
6. Multiple templates to be filled
7. Form-II requires classification of revenue, expenses, operating metrics, and margins.
8. Form-III requires asset, liability, capital structure, and working capital analysis.
9. Formula-based Excel used for further computation
10. Once forms are filled, excel sheets with complex formulas calculate ratios, cash flows, financial indicators, and projections.
11. Even small input mistakes can distort the entire analysis.
12. Limited automation → heavy dependence on manual effort
13. Significant time is spent validating numbers, matching totals, and cross-checking year-on-year figures.
14. Final credit report depends on manual accuracy
15. Output quality and reliability depend on the analyst’s skill and manual diligence, leading to inconsistent results and delays.

AI Solution

Current Solution – Key Highlights

1. Automated Extraction of Financial Statement (FS) Data

1. Upload multi-year annual reports (PDFs) directly into the system.
2. AI engine cleans, normalises, and maps financial data to predefined templates (Form-II & Form-III).
3. Eliminates manual copy-paste and reduces data-entry errors.
4. Delivers templated Excel files instantly to the user’s inbox.
5. Supports multi-entity and multi-year processing in a single upload.

2. AI-Driven Credit Analysis Report

1. Upload F-II and F-III Excel files to trigger real-time credit analysis.
2. System auto-generates a structured Credit Report JSON, ensuring consistent output.
3. Computes financial ratios, solvency metrics, liquidity scores, cash flow indicators, DSCR, covenants, and capex analysis.
4. Provides a multi-year metrics dashboard for trend analysis.
5. Enables export to Excel and download of ready-to-share HTML snapshots.
6. Offers a unified view of all credit indicators, removing dependency on manual formula sheets.

Technology Stack

Technology Used – Key Components

1. Backend & Workflow Automation – n8n

1. n8n orchestrates the entire backend workflow, from file ingestion to data extraction and processing.
2. Manages multi-step automation including PDF parsing, mapping logic, template generation, and report computation.
3. Ensures seamless integration between data sources, AI models, and output modules.

2. Backend Processing Engine

1. Custom backend logic prepares, normalises, and validates financial data.
2. Executes formula-based transformations, ratio calculations, and credit-scoring algorithms.
3. Provides consistent, structured outputs for both Excel and dashboard formats.

3. Frontend Interface

1. A user-friendly web interface designed for uploading files, triggering workflows, and downloading final reports.
2. Offers clean navigation, clear prompts, and interactive components for a smooth user experience.

4. HTML-Based Output Rendering

1. Uses HTML templates to generate downloadable credit snapshots and dashboards.
2. Enables visually rich, responsive, and shareable representations of financial and credit analysis.
3. Supports export to Excel, PDF, and JSON formats.

WORKFLOWS

1. REPORT INTAKE AND TRANSFORMATION CONSOLE

1. File Intake (Webhook Upload)
2. User uploads one or more annual report PDFs through the webhook/front-end.
3. n8n triggers the “Financial Report Extraction” workflow.
4. Split into Financial Sections
5. The workflow routes the PDF into multiple extraction branches:
6. Profit & Loss (P&L)
7. Balance Sheet (BS)
8. Notes to Accounts / Schedules (multi-branches)
9. AI Extraction for Each Block
10. For every branch, an AI/LLM node reads the relevant pages/sections.
11. Data is mapped to a structured schema (e.g., Revenue, COGS, EBITDA, Assets, Liabilities, Equity, etc.).
12. Write to Interim Tables
13. Each extracted block is written into a dedicated table / dataset (Set nodes).
14. Examples: BS – Current Assets, BS – Non-Current Assets, P&L – Income, P&L – COGS, P&L – Operating Expenses, Interest/Tax, etc.
15. Merge Structured Data
16. A central Merge node combines all extracted blocks into one unified dataset for the entity and year.
17. This ensures totals reconcile and all required line items are captured.
18. Generate Form-III (Balance Sheet Analysis)
19. Balance Sheet branches are stacked and aligned to the Form-III template.
20. Output is written into a prepared Excel skeleton for BS ratios and working-capital analysis.
21. Generate Form-II (Income Statement / Operating Statement)
22. P&L branches are stacked and aligned to the Form-II template.
23. Output is written into the Excel skeleton for margins, profitability and coverage ratios.
24. Combine F-II & F-III Packs
25. A second Merge node bundles the completed F-II and F-III sheets into one Excel file pack.
26. Store in Google Drive
27. The finished Excel pack is copied/saved into a designated Google Drive folder for record-keeping.
28. Trigger Next Processing (HTTP Request)
29. An HTTP Request node can call the Credit Analysis Report Tool or any downstream API, passing file links/JSON payload.
30. Download Link & Email to User
31. A “Download file” node prepares the final file.
32. A Gmail node sends an email to the user with the Excel pack attached / linked and a confirmation message.

This is your end-to-end Financial Report Extraction Tool: PDFs in → AI-mapped F-II & F-III Excel out → ready for automated credit analysis.

In this step, we include human intervention of verifying FII and FII Report downloaded in excel format and then we will upload those excel files in the next tool for final output with dashboard.

1. CREDIT ANALYSIS REPORT TOOL

Hard-coded RBA Tool – Step-Wise Workflow

(Input = Excel files generated by the previous Extraction Tool)

1. Upload Excel Files (Webhook Trigger)

1. User uploads the auto-generated F-II (P&L) and F-III (Balance Sheet) Excel files.
2. The workflow starts immediately through an n8n webhook.

2. Fan-Out Files (Routing Logic)

1. A JavaScript node reads the upload payload and identifies:
2. Which file is P&L (F-II)
3. Which file is Balance Sheet (F-III)
4. Files are routed to their respective extraction branches.

3. Extract P&L Sheet (Extract from XLSX)

1. The workflow extracts P&L metrics from the F-II Excel output of the previous tool.
2. Reads structured line items like:
3. Revenue, COGS, Gross Margin
4. Employee costs
5. Finance cost, Depreciation
6. EBITDA, EBIT, PAT

4. Clean + Standardise P&L Data (JavaScript)

1. Numbers are normalised (removal of commas, text-to-number).
2. Headings mapped to standard tags required by the RBA engine.
3. Ensures consistency for further processing.

5. Extract Balance Sheet (Extract from XLSX)

1. BS data is extracted from the F-III Excel output.
2. Reads line items like:
3. Current Assets / Liabilities
4. Net Worth
5. Borrowings
6. Fixed Assets
7. Working Capital items

6. Clean + Standardise BS Data (JavaScript)

1. Converts all values to comparable numeric formats.
2. Maps Balance Sheet items to internal schema (e.g., WC, TD, Tangible NW).
3. Handles missing values and creates placeholders.

7. Merge Both Sheets (Append Merge Node)

1. P&L dataset + Balance Sheet dataset merged into one complete Financial Dataset.
2. This creates a unified structure required for ratio and risk computations.

8. Pre-Processing & Derived Metrics (JavaScript)

1. Calculates derived values such as:
2. EBITDA Adjusted
3. Net Working Capital
4. Interest Coverage Inputs
5. Debt components
6. Ensures every ratio has clean and dependable inputs.

9. Run Hard-Coded Credit Analysis Engine (RBA Node)

1. Applies rule-based logic:
2. Liquidity Ratios (CR, Quick Ratio, WC Cycle)
3. Leverage Ratios (Debt/EBITDA, Debt/Equity, TNW tests)
4. Profitability Ratios (EBITDA %, PAT %, ROCE)
5. Coverage Ratios (DSCR, Interest Coverage)
6. Flags deviations, thresholds, breaches, and risk markers.
7. Generates a structured Credit Analysis JSON.

10. Apply Scoring Logic (JavaScript)

1. Converts ratio outputs into risk scores:
2. Green (Low Risk)
3. Yellow (Moderate Risk)
4. Red (High Risk)
5. Calculates the final RBA Score / Internal Grade.

11. Format Final Output (JavaScript)

1. Prepares the final RBA report structure:
2. Summary Table
3. P&L Overview
4. Balance Sheet Overview
5. Ratio Table
6. Risk Flags
7. Final Rating / Score
8. Output ready to be sent via:
9. API
10. Email
11. Front-end UI
12. Downloadable JSON/HTML

OUTPUT :