@axiom-lattice/examples-deep_research 1.0.26 → 1.0.28

package/src/agents/data_agent/skills/business-analytics/SKILL.md

@@ -0,0 +1,230 @@
+---
+name: business-analytics
+description: Generates comprehensive CFO-level business analysis reports from metrics data. Creates board-level reports with profit bridge analysis, cash conversion cycle diagnostics, GMROI metrics, strategic scenario modeling, and actionable executive summaries. Use when analyzing business performance data, generating financial reports, or conducting strategic business analysis.
+---
+
+# Business Analytics Report Generation
+
+## Overview
+
+This skill generates comprehensive, board-level business analysis reports similar to the "Confectionary Low-Margin Deep Analysis Report" style. Reports combine financial metrics, strategic insights, and actionable recommendations.
+
+## Report Structure
+
+Always organize reports using this hierarchy:
+
+```
+1. Executive Summary (1 page)
+   - Core conclusions (3-5 bullet points)
+   - Top 3 immediate actions
+   
+2. Model 1: Gross Margin & Channel Split
+   - Methodology note
+   - Summary data tables
+   - Channel breakdowns
+   
+3. Model 2: Profit Bridge
+   - Revenue to operating profit flow
+   - Key insight callouts
+   - Risk visualizations
+   
+4. Model 3: Cash Conversion Cycle + GMROI
+   - CCC calculations
+   - DSO structure breakdown
+   - GMROI analysis vs benchmarks
+   
+5. Model 4: Strategic Scenarios
+   - Option A/B comparisons
+   - P&L impact modeling
+   - Business model characteristics
+   
+6. Model 5: CFO/CEO Q&A
+   - 5 key questions answered
+   - Strategic conclusions
+   
+7. Action Roadmap
+   - Priority matrix (Impact vs Urgency)
+   
+8. Appendix
+   - Key financial metrics reference
+```
+
+## Writing Style Guidelines
+
+### Tone
+- Professional, board-level language
+- Direct and concise
+- Data-driven with clear sourcing
+- Action-oriented
+
+### Formatting
+- Use ASCII tables for data presentation
+- Include currency symbols (¥, $, €, SGD) consistently
+- Show percentages with 1 decimal place
+- Use visual bar charts with █ characters where helpful
+
+### Data Presentation
+- Always cite data sources
+- Include analysis period dates
+- Show calculations transparently
+- Flag methodology limitations
+
+## Key Report Components
+
+### Executive Summary Template
+
+```markdown
+## Executive Summary (1 page)
+
+### Core Conclusions
+
+1. [Key finding with metric]
+2. [Key finding with metric]
+3. [Key finding with metric]
+
+### Top 3 Immediate Actions
+1. [Action item] — [Expected impact]
+2. [Action item] — [Expected impact]
+3. [Action item] — [Expected impact]
+```
+
+### Profit Bridge Template
+
+```markdown
+## Model 2: Profit Bridge
+
+| Item | Amount | % of Revenue |
+|------|--------|--------------|
+| Revenue | [Amount] | 100.0% |
+| Less: COGS | [Amount] | [Percentage] |
+| **Gross Margin** | [Amount] | [Percentage] |
+| Less: [Expense 1] | [Amount] | [Percentage] |
+| Less: [Expense 2] | [Amount] | [Percentage] |
+| Plus: [Income item] | [Amount] | [Percentage] |
+| **Operating Profit** | [Amount] | [Percentage] |
+```
+
+### Cash Conversion Cycle Template
+
+```markdown
+## Model 3: Cash Conversion Cycle
+
+### CCC Components
+
+| Metric | Days | Calculation |
+|--------|------|-------------|
+| DIO (Days Inventory Outstanding) | [Days] | (Average Inventory / COGS) × 365 |
+| DSO (Days Sales Outstanding) | [Days] | (AR / Revenue) × 365 |
+| DPO (Days Payable Outstanding) | [Days] | (AP / COGS) × 365 |
+| **CCC** | **[Days]** | **DIO + DSO - DPO** |
+
+### Key Insight
+[Explanation of CCC implications]
+```
+
+### GMROI Template
+
+```markdown
+### GMROI Analysis
+
+| Metric | Value | Benchmark |
+|--------|-------|-----------|
+| Annual Gross Margin | [Amount] | — |
+| Average Inventory | [Amount] | — |
+| **GMROI** | **[Ratio]x** | 3.0–5.0x |
+
+### Interpretation
+[Analysis of GMROI vs benchmark]
+```
+
+### Strategic Scenario Template
+
+```markdown
+## Model 4: Strategic Scenarios
+
+### Scenario Comparison
+
+| Metric | Option A | Option B | Delta |
+|--------|----------|----------|-------|
+| Revenue | [Amount] | [Amount] | [Amount] |
+| Gross Margin | [Amount] | [Amount] | [Amount] |
+| Operating Profit | [Amount] | [Amount] | [Amount] |
+
+### Business Model Characteristics
+
+| Characteristic | Option A | Option B |
+|----------------|----------|----------|
+| [Attribute 1] | [Value] | [Value] |
+| [Attribute 2] | [Value] | [Value] |
+```
+
+### CFO/CEO Q&A Template
+
+```markdown
+## Model 5: CFO and CEO Decision Q&A
+
+### CFO Questions
+
+**Q1: [Question]?**
+[Data-driven answer with supporting analysis]
+
+**Q2: [Question]?**
+[Data-driven answer with supporting analysis]
+
+[Continue for 5 key questions]
+
+### CEO Questions
+
+**Q1: [Strategic question]?**
+[Strategic analysis and recommendation]
+
+[Continue for 5 strategic questions]
+```
+
+### Action Roadmap Template
+
+```markdown
+## Action Roadmap (Priority Matrix)
+
+### Priority Framework
+
+```
+Impact
+  High │  [Action 2]         [Action 1]
+       │  (Urgent/High)       (High/Planned)
+       │
+       │  [Action 3]         [Action 4]
+  Low  │  (Urgent/Low)        (Planned/Low)
+       └────────────────────────────────────
+              Urgent             Planned
+```
+
+### Immediate Actions (This Week)
+1. [Action item with owner and deadline]
+
+### Short-term Actions (This Month)
+1. [Action item with owner and deadline]
+
+### Medium-term Actions (This Quarter)
+1. [Action item with owner and deadline]
+```
+
+## Usage Workflow
+
+When generating a business analytics report:
+
+1. **Understand the data source** - Identify available metrics, tables, and time periods
+2. **Determine the analysis scope** - What business questions need answering?
+3. **Select relevant models** - Not all 5 models may be needed for every report
+4. **Calculate metrics** - Ensure all financial calculations are transparent
+5. **Write executive summary** - Lead with conclusions and actions
+6. **Build detailed models** - Follow the template structures
+7. **Create action roadmap** - End with clear, prioritized next steps
+
+## Examples
+
+See [examples.md](examples.md) for complete sample reports.
+
+## Reference
+
+For detailed methodology on specific calculations, see [reference.md](reference.md).

package/src/agents/data_agent/skills/business-analytics/examples.md

@@ -0,0 +1,295 @@
+# Business Analytics Report Examples
+
+## Example 1: Confectionery Business Analysis
+
+This example demonstrates a complete CFO-level business analysis report following the structure defined in SKILL.md.
+
+---
+
+# Candy (Confectionery) Low-Margin Deep Analysis Report v2
+## Candy Business Profit Restructuring Report (Board Level)
+
+**Data Source**: ZZZZ_KAIMAYMTC_TEST — MTC Enterprise, Southeast Asia F&B Distributor  
+**Analysis Period**: Full year 2025 (snapshot data as of 2026-03-03)  
+**Data Foundation**: OINV · OPCH · STOCK · BATCHSTOCK · CUSTBAL · JournalEntry (accounts 401/501 series)  
+**Report Version**: v2.1 — Profit Bridge Restructuring Edition; v2.1 Addendum: Cost/Pricing口径 Alert (§1.5), SKU Margin 77–99% Conclusion Withdrawn, P&L/JE为准
+
+---
+
+## Executive Summary (1 Page)
+
+### Core Conclusions
+
+The candy business is technically profitable (10.6% real operating margin) but structurally fragile due to three core constraints:
+
+1. **Subsidy Dependency Crisis**: Nestle subsidies contribute 42% of commercial gross profit; real operating profit drops to 1.1% if subsidies are lost
+2. **Bad Debt Time Bomb**: Accounts receivable aging at 256 days, with ¥2.9M (28%) overdue by more than 2 years, posing a threat of direct loss of 17% of net assets if fully provisioned at one time
+3. **Inventory Suspicion**: KitKat 2F 17g system shows 703,848 units, requiring 131 years to sell at current velocity; suspected unit of measure (case vs. piece) error, requiring immediate physical count verification
+
+### Top 3 Immediate Actions
+
+1. **Immediately**: Initiate export bad debt collection — Wei Jin customer (HLA + Indo China) ¥2.78M overdue by more than 2 years
+2. **This Week**: Dispose of Nutella & Go T1 (4,800 units, 4 days to expiry) + Emergency Ferrero T30 order (shortfall of 27,483 units)
+3. **This Month**: Renegotiate Nestle subsidy contract (¥1.71M/year, 46% of gross profit, extremely high renewal risk)
+
+---
+
+## Model 1: Gross Margin Methodology & Channel Split (JournalEntry Only口径)
+
+### 1.1 Methodology Note
+
+All gross profit and COGS in this report come from MTC_VW_AI_JournalEntry:
+
+- **Revenue**: Lines with account codes starting with 4, net amount = Credit − Debit (in base currency)
+- **COGS**: Lines with account codes starting with 5, net amount = Credit − Debit (usually negative, indicating cost incurred)
+- **Gross Margin** = (Revenue + COGS) / Revenue × 100% (since COGS is negative, numerator is revenue plus COGS)
+- **Candy category accounts**: Revenue 401020100, Cost 501030100 (see appendix for account mapping)
+- Can drill down by PrcName (CC1 Business Unit), PrcName2 (CC2 Channel), RefDate, etc.; PrcName coverage 100%, PrcName2 coverage ~96.5% in candy account lines (missing lines can be grouped into "Unassigned")
+
+**Note**: This report does not perform SKU-level gross profit estimation; all gross profit/COGS/gross margin conclusions are based on JE口径.
+
+### 1.2 Candy Business Summary (JE口径, 2025)
+
+| Metric | Value |
+|--------|-------|
+| Revenue | ¥19,228,590 |
+| COGS | -¥15,492,840 |
+| **Gross Margin** | **¥3,735,750** |
+| **Gross Margin %** | **19.4%** |
+
+### 1.3 Channel (CC2) Split
+
+For "Local vs Export" or specific channel gross profit, filter and aggregate candy accounts 401020100/501030100 net amounts (Credit−Debit) from JournalEntry by PrcName2 (and PrcName), ensuring revenue and COGS are from the same source and口径. COGS and gross profit in Model 2 (Profit Bridge) and Model 4 of this report are based on JE aggregate or JE channel split, with no non-JE口径 cost estimation.
+
+---
+
+## Model 2: Profit Bridge — From Theory to Real Operating Profit
+
+```
+                                          SGD        % of Revenue
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+Revenue (OINV Net)                    19,228,590     100.0%
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+Less: COGS (Historical Cost)           -15,492,840     -80.6%
+                                    ───────────   ───────
+[A] Book Gross Profit                   3,735,750     +19.4%
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+Less: Trading Terms (Promotions)           -502,145      -2.6%
+Less: Carriage Outwards (Delivery)       -420,798      -2.2%
+Less: Sales Discount                     -192,784      -1.0%
+Less: Expired Stock                      -122,889      -0.6%
+Plus: Nestle Subsidy + Commission      +1,829,032      +9.5%
+                                    ───────────   ───────
+[B] Commercial Gross Profit             4,326,166     +22.5%
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+Less: Bad Debt Provision (180d+ × 50%)   -2,175,000     -11.3%
+Less: Inventory Carrying Cost                                        
+     (¥1,338K × 8% Cost of Capital)       -107,000      -0.6%
+                                    ───────────   ───────
+[C] Real Operating Profit (CFO View)    2,044,166     +10.6%
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+```
+
+### Key Breakdown Insights
+
+**Subsidy Dependency Risk Visualization:**
+
+```
+Commercial Gross Profit 22.5%
+├── Subsidy Contribution:  +9.5%  ████████████████████████  (42% of Commercial GP)
+└── Core Operations:       13.0%  If subsidies disappear, commercial GP drops to 13.0%, real profit margin drops to ~1.1%
+```
+
+**CFO Core Warning:** The company's candy business is economically in "life support" mode via Nestle subsidies. Any subsidy negotiation failure will directly compress real profit to near break-even.
+
+---
+
+## Model 3: Cash Conversion Cycle (CCC) + GMROI — Capital Efficiency Diagnosis
+
+### 3.1 CCC Calculation
+
+| Metric | Days | Formula |
+|--------|------|---------|
+| DIO (Days Inventory Outstanding) | 31 | (Average Inventory / COGS) × 365 |
+| DSO (Days Sales Outstanding) | 256 | (AR / Revenue) × 365 |
+| DPO (Days Payable Outstanding) | 27 | (AP / COGS) × 365 |
+| **CCC** | **260** | **DIO + DSO - DPO** |
+
+**Inventory management is healthy (31 days), but AR severely drags (256 days), resulting in an overall CCC of 260 days.**
+
+This means the company has prepaid approximately 8.5 months of funds before receiving payment for this sale.
+
+### 3.2 DSO Structure Breakdown
+
+Among the 256-day DSO, export customers (AR ¥10.2M) are the main cause:
+
+```
+Export Customer AR ¥10.2M Distribution:
+  0–30 days (Normal):   ¥ 0.7M   7%   ██
+  31–90 days (Watch):   ¥ 3.5M  35%   ████████
+  91–180 days (Warning):¥ 1.0M  10%   ███
+  181–365 days (Overdue):¥0.9M   9%   ██
+  1–2 years (Bad debt risk):¥0.6M   6%   █
+  >2 years (Likely bad debt):¥2.9M  28%  ███████   ← Main Problem
+```
+
+43% of export AR is overdue by more than 6 months, with 28% overdue by more than 2 years, mostly concentrated in Wei Jin's Myanmar/Indochina customers.
+
+### 3.3 GMROI (Gross Margin Return on Inventory Investment)
+
+```
+GMROI = Annual Gross Margin ÷ Average Inventory Amount
+       = ¥3,735,750 ÷ ¥1,337,863
+       = 2.79x
+
+Industry Benchmark (FMCG Distribution): 3.0–5.0x
+MTC Current: 2.79x (Below benchmark minimum)
+```
+
+**Interpretation:** For every 1 yuan of inventory invested, 2.79 yuan of gross margin is generated  
+**FMCG Industry Standard:** Every 1 yuan of inventory generates 3–5 yuan gross margin
+
+**Reasons for low GMROI:**
+- Backlog inventory (Kopiko and other bulk commodities, not generating additional margin but occupying funds)
+- Near-expiry inventory risk (45,558 units expiring within 90 days)
+- High-cost goods (Ferrero, Guylian) not turning fast enough
+
+### 3.4 CCC Optimization Path
+
+| Initiative | Impact on CCC | Priority |
+|------------|---------------|----------|
+| Bad debt collection | -60 days | High |
+| AR factoring | -40 days | Medium |
+| Inventory optimization | -10 days | Medium |
+| Supplier payment terms | +5 days | Low |
+
+---
+
+## Model 4: Strategic Focus Simulation — Cut Local vs Expand Export Real P&L Impact
+
+### 4.1 Scenario Comparison
+
+| Metric | Cut Local (Scenario A) | Expand Export (Scenario B) | Delta |
+|--------|------------------------|---------------------------|-------|
+| Revenue | ¥17,033,590 | ¥24,719,385 | +¥7,685,795 |
+| Gross Margin | ¥3,296,000 | ¥5,257,000 | +¥1,961,000 |
+| Operating Profit | ¥1,890,000 | ¥2,890,000 | +¥1,000,000 |
+| Profit Margin | 11.1% | 11.7% | +0.6% |
+
+### 4.2 True Characteristics of Both Business Models
+
+```
+                Export Brand Distribution    Local FMCG Distribution
+                ─────────────────────────    ─────────────────────
+Customer Count          8–42                        200–300+
+Avg Invoice            ¥51K–63K                    ¥2.5K–8K
+Unit Gross Margin      *94–97% (口径存疑)          *77–97% (口径存疑)
+Absolute Margin/Unit   ¥115–260                    ¥15–85
+Delivery Cost          Low (Container)             High (Fragmented, Daily/Weekly)
+TT Share               High (Brand needs export)   Medium (Local listing fee)
+AR Period              Long (Avg 256 days)         Short (30–60 days)
+Bad Debt Risk          Extreme (Wei Jin case)      Minimal
+Subsidy Dependency     High                        Medium
+Cash Recovery          Poor                        Good
+
+Core Paradox: Export high unit price creates high margin, but bad debt and
+            capital cost almost completely offset
+            Local low unit price but high efficiency, true cash flow engine
+
+Strategic Conclusion: The two models are not mutually exclusive but complementary —
+                     Local business is the cash flow engine, export is the profit scale engine.
+                     The problem is not choosing which, but export must implement strict credit management.
+```
+
+---
+
+## Model 5: CFO and CEO Decision Q&A
+
+### CFO's 5 Core Questions
+
+**Q1: If local business is cut, will profit rise or fall?**
+
+Needs recalculation with data. Previous "local SKU-level margin 77–99%" based on STOCK.AvgPrice and OINV Price has cost/pricing口径 issues (§1.5), not reliable. Should first use JE channel-split COGS and revenue to calculate true local channel marginal contribution, then assess gross margin change and operational savings after cutting; cannot assert "must lose money" or "must improve" before completing this calculation.
+
+**Q2: If Nestle subsidies reduce by 50%, can we survive?**
+
+Barely. 50% subsidy reduction = ¥850K loss, real operating profit drops from ¥2.05M to ~¥1.2M (6.2%), still positive but extremely fragile. If bad debt worsens simultaneously, then losses. Must immediately initiate multi-brand subsidy agreements (Ferrero, Lotte) to diversify risk.
+
+**Q3: If bad debt is fully provisioned at once, what's the net asset impact?**
+
+Bad debt ¥4.345M at 80% actual loss rate = ¥3.476M (SGD) loss. If total net assets are ~¥20M, one-time recognition would erode ~17%. Need to provision in batches over 2–3 years to smooth impact.
+
+**Q4: Is export customer concentration controllable?**
+
+Not controllable, and high risk. Wei Jin alone manages Myanmar/Indochina customer base, contributing ~38% of export AR, with these customers accumulating ¥2.78M bad debt. Single salesperson concentrating management of cross-border high-risk customers is a major governance defect. Recommend immediately establishing multi-person cross-coverage and direct company-to-customer contract signing mechanisms.
+
+**Q5: Is inventory book-to-physical consistent?**
+
+Major doubts exist. KitKat 2F 17g system shows 703,848 units, requiring 131 years to sell at current velocity. Kopiko series multiple SKUs at 66,000–79,000 units each, requiring 13–47 years to sell. Highly suspected unit of measure (case vs piece) error, requiring immediate physical count verification. If true, asset misstatement may affect bank credit assessment.
+
+### CEO's 5 Strategic Questions
+
+**Q1: What business should the company actually be in?**
+
+Do both, but must clearly separate governance:
+- Export business = profit growth engine, needs strict credit policy protection
+- Local business = cash flow engine, needs MOQ policy to reduce operating costs
+
+Cannot manage them together, need to establish independent P&L tracking.
+
+**Q2: What if Desmond Tan leaves?**
+
+He alone contributes ¥6.14M (31% of candy revenue), serving 8 major export customers. This is unacceptable key person risk. Immediately initiate: dual customer relationship coverage, direct company-to-customer contract signing, Desmond customer information systematization and archiving.
+
+**Q3: What happens when Lotte Pepero IP co-branded popularity fades?**
+
+Pepero series (Stray Kids/New Jeans) accounts for ~18% of 2025 candy revenue. IP popularity cycle ~6–12 months. Need to establish: IP co-branded purchase volume control (max 6 months inventory), while introducing non-IP standard version (Pepero standard) as long-term stable product.
+
+**Q4: Can export expand without increasing bad debt?**
+
+Yes, but needs conditions: ① New export customers must use LC (letter of credit) or prepayment; ② Existing Wei Jin customers set credit freeze (no additional exposure); ③ Internal per-customer maximum credit limit (suggest ¥1M export customer single limit). Under this framework, ¥5.49M export expansion can add ¥5.21M profit.
+
+**Q5: If nothing changes, what happens in 3 years?**
+
+- Subsidy renewal failure probability increases over time
+- Bad debt balance accumulates annually (current ¥4.34M → if another 30% then exceeds ¥5.65M)
+- Inventory backlog continues (tying up capital, increasing expiry losses)
+- CCC 260-day structural problem drags capital efficiency
+
+High probability of cash flow crisis or need for shareholder capital increase within 3 years. Must initiate structural reform within next 6 months.
+
+---
+
+## Action Roadmap (Priority Matrix)
+
+```
+Impact
+  High │  ② Bad Debt Collection      ① Credit Policy Reform
+       │  (Urgent/High Impact)       (High Impact/Plannable)
+       │
+       │  ⑤ Subsidy Renewal          ④ Export Scale Expansion
+  Mid  │  (Urgent/Mid Impact)        (Mid-term/High Impact)
+       │
+       │  ③ Inventory Disposal       ⑥ Local MOQ Policy
+  Low  │  (Must Do This Week)         (Mid-term/Mid Impact)
+       └─────────────────────────────────────────────────────
+              Urgent                   Plannable
+```
+
+## Appendix: Key Financial Metrics Quick Reference
+
+| Metric | Formula | Benchmark | Current |
+|--------|---------|-----------|---------|
+| Gross Margin % | (Revenue - COGS) / Revenue | 15-25% | 19.4% |
+| Operating Margin % | Operating Profit / Revenue | 8-15% | 10.6% |
+| DSO | (AR / Revenue) × 365 | 30-60 days | 256 days |
+| DIO | (Inventory / COGS) × 365 | 20-40 days | 31 days |
+| CCC | DIO + DSO - DPO | 60-100 days | 260 days |
+| GMROI | Gross Margin / Avg Inventory | 3.0-5.0x | 2.79x |
+
+---
+
+**Report Version**: v2.1 (Board-Level Profit Restructuring Edition; includes Cost/Pricing口径 Correction Note)  
+**Generated**: 2026-03-03  
+**Data Source**: ZZZZ_KAIMAYMTC_TEST / SAP B1 HANA  
+**Analysis Dimensions**: Financial Profit Bridge · SKU×Channel Matrix (口径存疑) · Cash Conversion Cycle · Strategic Simulation · Decision Q&A