📈 Stock Analysis Agent

5-agent fundamental deep-read pipeline — demonstrated on 99 Speed Mart (99SMART) Q3 2025 interim financial report

Bursa Malaysia Gemini 2.0 Flash 5-Agent Pipeline Interim Report Deep-Read Investment Research
🚀 Launch AI Assistant Demo 📁 View Source Code

📋 Project Overview & Problem Statement

Challenge: A Malaysian retail investor opens 99 Speed Mart's Q3 2025 interim report — 20 pages, thousands of numbers, five separate statements. Reading it cover-to-cover takes hours, and most readers still come away unable to answer the three questions that matter: is this a quality business, is the margin trend real, and is the dividend safe to bank on? Missing the signal buried in the filing is the single most expensive thing a retail investor does.

Solution: The Stock Analysis Agent runs a Bursa-listed stock's interim filing through a 5-agent AI deep-read pipeline — screening, fundamentals, moat, dividends, and reporting. This showcase demonstrates the pipeline applied to 99 Speed Mart (Bursa: 99SMART) Q3 2025. Output: screening verdict, margin trend read, moat-widening inventory, dividend-sustainability analysis, and a structured executive summary with top investment merits and top risks — in minutes, not hours.

Key Benefits

🤖 Multi-Agent AI Architecture

🔍 Agent 1: Screening Agent

Runs the basic health checks — growth rate, profitability, EPS, net assets per share. In the 99SMART case: flagged PASS on both top-line growth (+19.1% Q3 revenue) and profitability (all margins expanding).

📈 Agent 2: Fundamentals Agent

Reads margin trends, cost discipline, and operating leverage from the P&L. In the 99SMART case: revenue +19.1% while opex only +12.6% — textbook operating leverage, PBT margin expanded +1.3 pp.

🏰 Agent 3: Moat Agent

Assesses scale, geographic reach, and moat-widening initiatives. In the 99SMART case: 2,966 outlets across all Malaysian states, first China outlet in Fuzhou, bulk-sales e-commerce adding incremental revenue — four separate moat moves live at once.

💰 Agent 4: Dividends Agent

Tracks payout record, coverage by operating cash flow, and balance-sheet strength. In the 99SMART case: RM378M FY2025 dividends declared (3.8× FY2024) against RM787M operating cash — 2.08× coverage. Cash pile RM1.07B, term loans zero.

📋 Agent 5: Reporting Agent

Synthesises the four upstream agents into an executive summary, top 3 investment merits, and top 3 risks. In the 99SMART case: growth-plus-income story with operating leverage as merit #1, payout ratio (82.7%) as risk #1.

Investment Yardstick — the analytical rubric the pipeline applies to every stock

The Yardstick below is the 5-agent pipeline's full analytical lens. For the 99SMART demo, the pipeline produces qualitative findings against each category (covered in the Launch Demo above). The same Yardstick also drives the optional batch-screener mode for 50+ Malaysian stocks via Yahoo Finance.

Category Weight Max Points Evaluation Criteria
Fundamentals & Profitability 20% 20 ROE, EPS, profit margins, operating efficiency
Financial Health & Solvency 15% 15 Debt ratios, current ratio, book value, liquidity
Valuation 20% 20 PE ratio, price-to-book, price-to-sales
Business Moat 15% 15 Competitive advantages, brand, barriers
Cash Flow & Dividends 15% 15 Dividend yield, payout ratio, sustainability
Management & Outlook 10% 10 Track record, governance, strategy
Liquidity 5% 5 Trading volume, bid-ask spread

Recommendation Scale

🛠️ Technical Architecture & Data Sources

AI & Processing Stack

Google Gemini 2.0 Flash Yahoo Finance API Tavily Web Search Python 3.8+ Google Colab

Data Analysis Libraries

Pandas yfinance Matplotlib Seaborn JSON Processing

Multi-Agent Workflow — as applied to the 99SMART case

Sequential Agent Execution:

Data Sources & Integration

📖 Usage Guide & Examples

Analyse a Single Stock — 99SMART example

# Deep-read the 99 Speed Mart Q3 2025 interim report results = run_stock_analysis_pipeline( ticker='5326.KL', report_pdf='99SMART_Q3_2025_interim.pdf' ) # View generated report and visualisation print(f"Report: {results['report_path']}") print(f"Chart: {results['chart_path']}") # Opens markdown report with: # - Screening verdict (top-line growth + profitability checks) # - Fundamentals read (margin trends, operating leverage) # - Moat inventory (scale, geographic reach, new channels) # - Dividend sustainability (payout ratio, cash coverage) # - Executive summary + top 3 merits + top 3 risks

Stock Discovery (Batch Screening)

# Screen 50 Malaysian stocks and analyze top 3 discovery_results = discover_stocks( criteria={ 'max_price': 1.00, # Price ≤ RM 1.00 'min_pe': 4, 'max_pe': 15, # PE ratio 4-15 'min_roe': 5, # ROE ≥ 5% 'min_dividend_yield': 1, # Dividend yield ≥ 1% 'min_market_cap': 50_000_000, # Market cap ≥ RM50M 'min_volume': 50_000 # Volume ≥ 50K/day }, max_stocks_to_scan=50, analyze_top_n=3 ) # View passing stocks print(discovery_results['passing_stocks'])

Supported Malaysian Stocks

50+ Major Stocks Including:

🚀 Getting Started with Google Colab

Quick Start (3 Steps)

Example Analysis Output — 99SMART Q3 2025

Investment Report Contents:

📊 Performance Metrics & Capabilities

50+
Stocks Screened
5
Specialized AI Agents
7
Analysis Categories
100
Point Scoring Scale

Analysis Speed & Efficiency

Investment Criteria Supported

⚠️ Important Disclaimer

FOR EDUCATIONAL PURPOSES ONLY

This tool is designed for educational and research purposes. It is NOT financial advice. Stock market investing involves risk, and you could lose money.

User Responsibilities

The AI agents provide analytical insights based on publicly available data, but investment decisions should always be made with professional guidance and personal risk assessment.

🚀 Try the AI Assistant Demo 📖 Full Documentation