AI Skill Report Card
Multi Perspective Analysis
Multi-Perspective Analysis
Quick Start15 / 15
Input: "Should we launch this AI product in the EU market?"
Immediate Analysis:
- Legal: GDPR compliance gaps identified, 6-month certification needed
- Market: €2.3B addressable market, 67% adoption probability
- Risk: Regulatory changes likely Q2 2024, competitive threats from 3 incumbents
- Technical: Architecture supports GDPR by design, 99.7% uptime requirement
- Financial: €1.2M investment, 18-month break-even, 34% IRR
- Synthesis: Proceed with staged launch in Germany first, then EU-wide
Recommendation▾
Consider adding a template or framework section with a standardized format for conducting the analysis
Workflow14 / 15
Step 1: Lens Selection
- Identify 3-5 relevant analytical perspectives
- Map each lens to specific evaluation criteria
- Assign confidence weighting based on data availability
Step 2: Parallel Analysis
- Legal: Compliance requirements, regulatory risks, liability exposure
- Financial: Cost-benefit analysis, ROI calculations, cash flow impact
- Technical: Feasibility assessment, implementation complexity, scalability
- Market: Customer demand, competitive landscape, positioning opportunities
- Risk: Threat modeling, mitigation strategies, scenario planning
Step 3: Cross-Validation
- Compare findings for consistency and conflicts
- Identify assumptions requiring validation
- Flag high-uncertainty areas for deeper investigation
Step 4: Synthesis
- Weight recommendations by expertise relevance and data quality
- Create integrated decision framework
- Develop implementation roadmap with checkpoints
Recommendation▾
Include time estimates for each workflow step to help users manage analysis scope
Examples20 / 20
Example 1: Software Architecture Decision Input: "Microservices vs monolith for our 50-person engineering team?" Output:
- Technical: Monolith better for current team size, microservices add 40% complexity overhead
- Operational: Current DevOps maturity insufficient for microservices (Level 2/5)
- Financial: Monolith saves $180K annually in infrastructure and tooling costs
- Risk: Monolith creates single point of failure but reduces operational risk by 60%
- Recommendation: Start monolith, plan microservices transition at 100+ engineers
Example 2: Market Entry Strategy Input: "Enter Indonesian fintech market with digital payments app?" Output:
- Regulatory: Bank Indonesia requires local partnership, 18-month licensing process
- Cultural: Cash preference 78%, mobile-first adoption accelerating 23% annually
- Competitive: 12 established players, consolidation phase beginning
- Financial: $3.2M minimum viable entry, break-even at 500K active users
- Recommendation: Partner with local bank, focus on rural markets (underserved)
Recommendation▾
Add guidance on when NOT to use this approach (simple decisions, time-critical situations)
Best Practices
Lens Prioritization
- Weight analysis by relevance: technical depth for engineering decisions, market analysis for product launches
- Use 80/20 rule: focus 80% effort on most critical 2-3 perspectives
- Cross-check assumptions between complementary lenses (legal + financial for compliance costs)
Data Integration
- Quantify qualitative insights where possible (regulatory risk = probability × impact)
- Use common metrics across lenses (time to implementation, cost implications)
- Flag data conflicts early and investigate root causes
Decision Framework
- Create go/no-go criteria before analysis begins
- Build decision trees for complex scenarios
- Include sensitivity analysis for key assumptions
Common Pitfalls
Analysis Paralysis
- Don't analyze every possible angle—limit to 5 perspectives maximum
- Set analysis timeboxes: 2 hours per lens for most decisions
- Use "good enough" threshold rather than perfection
Confirmation Bias
- Actively seek disconfirming evidence in each analytical lens
- Assign one lens to play "devil's advocate" role explicitly
- Weight contrary evidence appropriately rather than dismissing it
False Precision
- Don't over-engineer financial models with dozens of variables
- Present ranges rather than point estimates for uncertain metrics
- Acknowledge when data quality is insufficient for confident conclusions
Integration Failures
- Avoid simply listing separate analyses—synthesize into unified recommendation
- Resolve contradictions between lenses rather than presenting conflicting advice
- Ensure final recommendation addresses insights from all relevant perspectives