--- name: ai-strategy-report argument-hint: "[company name] [industry] [focus area] e.g. ABC Manufacturing, Auto Parts, Cost Reduction" description: "Generate comprehensive AI strategy reports (15+ slides) in PowerPoint format. Analyzes company's digital readiness, identifies high-value AI use cases, creates implementation roadmaps, and provides ROI projections. Designed for consulting firm quality deliverables." allowed-tools: "Read Write Bash" metadata: version: 1.0.0 description_zh: "生成完整的AI战略报告PPT（15+页），分析企业数字化现状、识别高价值AI场景、制定实施路线图、提供ROI预测。咨询公司级别的交付物。" category: digital-transformation ---

# AI Strategy Report

## Overview

AI Strategy Report is a comprehensive strategic document that analyzes a company's AI readiness, identifies high-value use cases, and creates a practical implementation roadmap. This skill generates **professional-grade PowerPoint presentations of 15+ slides** using the **KPMG consulting template** for consistent, professional formatting.

**Key Features:** - **Complete strategic framework**: 15-slide structure covering the full AI transformation journey - **Data-driven analysis**: Digital maturity assessment and data readiness evaluation - **Prioritization matrix**: 2×2 value-feasibility matrix for AI opportunities - **Implementation roadmap**: 3-phase timeline (0-6m/6-18m/18-36m) - **Financial projections**: Investment breakdown and ROI calculations - **Risk assessment**: Technical, organizational, and compliance risks with mitigations

**Output Format:** PowerPoint (.pptx) with professional styling and structured content.

## When to Use This Skill

This skill should be used when: - Developing AI transformation strategy for enterprises - Evaluating AI opportunities and prioritizing use cases - Creating implementation roadmaps for digital transformation - Building business cases for AI investments - Assessing organizational readiness for AI adoption - Planning talent and capability building for AI teams - Preparing board-level presentations on AI strategy - Supporting M&A due diligence for AI capabilities - Planning cloud migration and data infrastructure - Creating vendor selection criteria for AI platforms

## Report Structure (15 Slides)

``` Slide 1: Cover Page Slide 2: Executive Summary Slide 3-4: Current State Assessment Slide 5-6: AI Opportunity Map (2×2 Matrix) Slide 7-9: Top 3 Use Cases Deep Dive Slide 10-11: Implementation Roadmap (3 Phases) Slide 12: Investment & ROI Analysis Slide 13: Organizational Capabilities Slide 14: Risk Assessment & Mitigation Slide 15: Next Steps & Action Items ```

## Input Requirements

### Required Information

```markdown **Company Basics** - Company Name: [Name] - Industry: [Industry Sector] - Company Size: [Employees] / [Revenue] - Digital Maturity: [Beginner/Intermediate/Advanced]

**Business Context** - Core Business: [Description] - Key Challenges: [List 2-3 major pain points] - AI Objectives: [What problems to solve with AI]

**Strategic Priorities** (Select all that apply) - [ ] Cost Reduction & Efficiency - [ ] Revenue Growth - [ ] Customer Experience - [ ] Innovation & New Products - [ ] Risk Management ```

### Optional Information

```markdown **Data Assets** - Existing data types: [Customer/Operational/IoT/etc] - Data history: [Years of historical data]

**Technology Stack** - Cloud platform: [AWS/Azure/GCP/Alibaba/etc] - Existing systems: [ERP/CRM/MES/etc]

**Constraints** - Budget range: [Amount] - Timeline: [Expected delivery] - Special restrictions: [Data privacy/etc] ```

## Workflow

### Phase 1: Analysis (Internal)

Analyze the input information and determine: 1. **Digital maturity level** based on described systems and processes 2. **Data readiness** for each potential AI use case 3. **Priority ranking** of AI opportunities (value × feasibility) 4. **Implementation complexity** for each phase

### Phase 2: Content Generation

Generate structured content for each slide:

**Slide 2 - Executive Summary:** - 3-5 key conclusions - Investment overview - Expected ROI - Critical milestones

**Slide 5-6 - AI Opportunity Map:** Create a 2×2 matrix categorizing opportunities: - **Quick Wins** (High Value, High Feasibility): Immediate start - **Strategic Bets** (High Value, Low Feasibility): Long-term planning - **Low Priority** (Low Value): Defer or discard

**Slide 10-11 - Roadmap:** Define 3 phases: - **Phase 1 (0-6 months)**: Foundation + Pilot - **Phase 2 (6-18 months)**: Scale + Capability Building - **Phase 3 (18-36 months)**: Optimization + Innovation

### Phase 3: Tool Execution

Call `generate_ppt_from_skill` tool with structured slide content:

```json { "skill_name": "ai-strategy-report", "title": "[Company] AI Strategy Report", "subtitle": "Digital Transformation Roadmap", "slides": [ { "type": "title", "title": "Cover Title", "content": "Subtitle" }, { "type": "content", "title": "Slide Title", "content": "Bullet points and analysis" }, { "type": "two_column", "title": "Comparison Slide", "left_content": "Current State", "right_content": "Future State" } ] } ```

### Phase 4: Optional Data Export

If user needs editable data, call `save_json`:

```json { "filename": "[Company]_AI_Strategy_Data", "data": { "scenarios": [...], "roadmap": {...}, "financial": {...} } } ```

## Tool Configuration

### Tool 1: generate_ppt_from_skill

**Purpose**: Generate PowerPoint using the KPMG template bundled with this skill

**When to Call**: After content generation is complete, always call this tool to create the deliverable.

**Parameters**: ```json { "skill_name": "ai-strategy-report", "title": "Company AI Strategy Report", "subtitle": "Digital Transformation Roadmap (2024-2027)", "slides": [ { "type": "title|content|two_column", "title": "Action-oriented title (verb-first)", "content": "Markdown formatted content with bullet points", "left_content": "For two-column layout", "right_content": "For two-column layout" } ] } ```

**Content Guidelines**: - Use action-oriented titles ("Drive Efficiency Through AI-Powered Quality Control") - Format with Markdown: `- Bullet points`, `**Bold highlights**` - Keep bullet points concise (1-2 lines each) - Use color coding: 🔴 High Risk, 🟡 Medium Risk, 🟢 Low Risk

### Tool 2: save_json (Optional)

**Purpose**: Export structured data for further editing or integration

**When to Call**: When user explicitly asks for editable data or mentions integrating with other systems.

**Parameters**: ```json { "filename": "Company_AI_Strategy_Data", "data": { "company": "Company Name", "industry": "Industry Sector", "scenarios": [...], "roadmap": {...}, "financial": {...}, "organization": {...}, "risks": [...] } } ```

## Quality Standards

### Content Requirements

- **Specificity**: All recommendations must be specific to the company's industry and stated challenges - **Quantification**: Include estimated savings/returns where possible (mark as "estimated" if not precise) - **Feasibility**: Only recommend AI use cases that match the described data availability - **Actionability**: Every recommendation must have clear next steps

### Slide Content Standards

**Executive Summary (Slide 2)**: - Max 5 conclusions - Include 1-line ROI summary - List 3 critical milestones

**Opportunity Map (Slide 5-6)**: - Minimum 4 opportunities mapped - Clear rationale for each quadrant placement - Prioritization within each quadrant

**Use Case Deep Dive (Slide 7-9)**: For each of top 3 use cases: - Business pain point (2-3 sentences) - AI solution approach (high-level) - Quantified expected benefit - Implementation complexity rating

**Roadmap (Slide 10-11)**: - Each phase has clear deliverables - Logical dependencies between phases - Resource requirements specified

**ROI Analysis (Slide 12)**: - 3-year investment breakdown - Year-by-year savings projection - Payback period calculation - Key assumptions listed

### Prohibited Content

- Do NOT specify specific vendors (e.g., "use AWS SageMaker") - Do NOT make unrealistic claims (e.g., "100% automation") - Do NOT ignore stated constraints (e.g., data privacy requirements) - Do NOT provide implementation details beyond strategic level

## Example Output

See `examples/manufacturing_example.md` for a complete input-output example.

## Best Practices

### For High-Quality Output

1. **Encourage detailed input**: If user input is vague, ask clarifying questions 2. **Be conservative with estimates**: Better to under-promise than over-promise 3. **Highlight risks explicitly**: Don't hide implementation challenges 4. **Emphasize data readiness**: Make clear when data preparation is needed 5. **Provide alternatives**: Offer options when ideal path isn't feasible

### Industry Customization

**Manufacturing**: - Focus: Predictive maintenance, quality control, supply chain - Key metrics: OEE, defect rates, inventory turnover

**Retail/E-commerce**: - Focus: Demand forecasting, personalization, pricing - Key metrics: Conversion rate, customer LTV, inventory accuracy

**Financial Services**: - Focus: Risk modeling, fraud detection, customer service - Key metrics: False positive rate, processing time, compliance score

**Healthcare**: - Focus: Diagnostic imaging, patient triage, resource optimization - Key metrics: Diagnostic accuracy, wait times, resource utilization

## Dependencies

### Required Backend Tools - `generate_ppt` - python-pptx 1.0.2 - `save_json` - Python built-in json

### System Requirements - AriaAI Backend >= 1.0.0 - Function Calling support enabled

## Version History

| Version | Date | Changes | |---------|------|---------| | 1.0.0 | 2024-03-25 | Initial release |

## Maintenance

- **Maintainer**: AriaAI Team - **Update Cycle**: Quarterly review - **Feedback**: Submit via Issue or contact admin

## Capability Upgrade

### Mode Selection

- **Quick**: 输出 AI 机会清单、优先级和 90 天试点建议。 - **Standard**: 输出完整 AI 战略报告、用例组合、路线图、投资和组织能力建议。 - **Deep**: 结合客户行业、数据资产、系统架构、组织成熟度、知识库案例和历史项目记忆，形成董事会级 AI 转型方案。

### AI Portfolio Logic

每个 AI 用例必须同时评估：业务价值、数据可得性、技术可行性、组织准备度、风险合规、落地周期和可复制性。优先级不能只按“看起来先进”排序。

### Quality Gates

- [ ] AI 用例与客户业务痛点和数据资产匹配。 - [ ] 投资收益有假设、区间和验证方式。 - [ ] 路线图区分数据基础、模型能力、业务流程和组织变革。 - [ ] 风险覆盖数据隐私、模型准确性、合规、采纳和运维。 - [ ] PPT 输出前已有清晰 storyline，不直接堆幻灯片。

## Consulting Excellence Layer

### AI Value Pool Logic

AI strategy must quantify value pools before listing use cases. Organize value into:

| Value Pool | Typical Levers | Evidence Needed | |------------|----------------|-----------------| | Revenue growth | Conversion, pricing, cross-sell, retention | Funnel, customer, sales data | | Cost reduction | Automation, workload reduction, rework reduction | Process volume, FTE, cycle time | | Risk control | Fraud, compliance, quality, safety | Incidents, exceptions, loss data | | Decision quality | Forecasting, planning, prioritization | Historical decisions and outcomes | | Knowledge leverage | Proposal reuse, case retrieval, expert assistance | Document corpus and usage patterns |

### Use Case Investment Committee

Every AI use case must be described as an investment case:

- Business problem. - User and workflow. - Data required. - Model approach. - Integration point. - Human review point. - Benefit hypothesis. - Risk and control. - Pilot metric. - Scale condition.

### Build / Buy / Partner Decision

| Condition | Recommended Path | |-----------|------------------| | Commodity capability, low differentiation | Buy SaaS or API | | Proprietary data and workflow advantage | Build on internal data | | Need speed plus domain expertise | Partner / co-build | | High compliance or sensitive data | Private deployment or controlled harness |

### AI Governance Minimum

Deep AI strategy must include:

- Model ownership and approval. - Data access and permission rules. - Prompt and output review policy. - Evaluation metrics and regression