Domain II: Identify Business Needs and Solutions — Pointer

Status: Pointer to existing material

Per the locked decision in CLAUDE.md, Domain II does not get a new study guide. scored Target on first attempt on this domain in the first attempt and the existing format-reference document — source/CPMAI_Phase_I_Study_Guide.md (52 lessons, 9 modules, ~929 lines) — already covers Phase I / Domain II content comprehensively.

• 📖 source/CPMAI_Phase_I_Study_Guide.md — comprehensive Phase I guide

• 📖 docs/ECO_TASK_REFERENCE.md — Domain II section (10 tasks, II.1 through II.10)

Domain II ECO Tasks (verbatim from docs/ECO_TASK_REFERENCE.md)

What to study for Domain II

Existing Phase I guide covers this comprehensively. Specifically, these sections are the most directly tested in Domain II:

• The Loop (Three Ps of Intelligence) — Lesson 2 of Phase I guide
• Probabilistic vs Deterministic Systems — Lesson 3
• Pyramid — Strategic Questions section
• Seven Patterns of AI — Pattern matching is heavy in II.1, II.2, II.7
• Go/No-Go Traffic Light Framework (9 lights, 3 categories) — supports II.2 (feasibility), II.6 (risk), II.10 (resources)
• 5 Strategic Questions (WHY AI? HOW fit? HOW build? HOW succeed?)
• ,,, mnemonics — (already memorized in Phase I)
• Trustworthy AI in Phase I (Lesson 38: 5 categories of principles, 3 transparency types) — pulls into Domain I

What's new for Domain II in the rebuild

Two refresh-only items that complement the existing Phase I guide:

### 1. PM Oversight Angle awareness (cross-domain reinforcement)

Even though scored Target on first attempt on Domain II, the PM Oversight Angle pattern he learned for Domains III/IV/V should be applied retroactively to Domain II questions on the retake. Every Domain II task starts with an oversight verb: identify, evaluate, conduct, develop, determine, manage, draft, define, support, identify. Same framing.

### 2. ROI and Business Case framing (II.5 + II.9)

Specific points to reinforce:

• ROI for AI projects ≠ ROI for traditional projects. AI value often comes from probabilistic-decision improvement, not deterministic cost savings. Account for: error reduction value, decision-speed value, scalability value (humans can't review every customer; AI can).
• Business case = stakeholder-engaged justification. PM coordinates the case; doesn't draft alone.
• ROI calculations include trustworthy-AI cost. Audit trails, bias monitoring, governance, contestability — all have cost. Include in case.

### 3. Success Criteria (II.8) feeds every other domain

Success criteria from II.8 are the single most cross-pulled artifact in the exam. They flow into:

• Domain III evaluation (III.7) and gate (III.8) — data is evaluated against success criteria.
• Domain IV gate (IV.6) — model is verified against success criteria.
• Domain V metrics (V.4) — production metrics tie to success criteria.

A weak success-criteria document compounds problems across all later phases. The PM facilitates rigorous definition in Phase I — measurable, time-bound, business-value-tied.

Cross-Domain Links (from Domain II)

• II.1 (Problem) ↔ I.3, I.4: Problem definition includes ethical and regulatory risk assessment.
• II.2 (AI Feasibility) ↔ III.1 (Required Data): Feasibility assessment includes "do we have the data?" → flows to III.1.
• II.3 (Risk Assessment) ↔ I.1, I.3, I.4, I.5 + V.7: Risk includes security, ethics, compliance, accountability + contingency.
• II.7 (Draft AI Solution) ↔ IV.1 (Technique): Draft solution constrains technique selection in Phase IV.
• II.8 (Success Criteria) ↔ Every other domain. The single most cross-pulled artifact.
• II.10 (Project Resources) ↔ III.4, V.1: Resources flow into infrastructure coordination and deployment planning.

Knowledge Check (Domain II refresh)

Use the existing Phase I guide's Knowledge Check section (lines ~875-905 of source/CPMAI_Phase_I_Study_Guide.md). Specifically these questions test cross-domain awareness:

• "During a Go/No-Go assessment, is it OK to move forward with red lights?" (False) — flows into III.8, IV.5, IV.6 gate logic.
• "Phase I asks 'What problem are we solving?'" (True) — every Domain II answer connects to this.
• "What are the repercussions of an AI system not providing clear information about its design?" — Domain I (transparency) pulled into Domain II.

Pitfalls Specific to Domain II

• AI for AI's sake. Wrong-answer trap: project where the technical capability exists but the business problem isn't well-defined. Right answer always returns to "what problem are we solving?" (II.1).
• Vague success criteria. Success criteria like "improve customer satisfaction" without measurable targets are PMI's red flag. Right answer: facilitate stakeholder engagement to define measurable, time-bound criteria (II.8).
• Underestimated trustworthy-AI cost. ROI calculations that omit governance/audit/bias costs are insufficient. Right answer: include Domain I costs in II.5 / II.9.
• Resource gaps surfaced too late. II.10 (resources) should identify gaps in Phase I, not Phase IV when training is blocked.

Closing reminders for Domain II

• scored Target on first attempt — don't over-invest here. Refresh, then focus weak-domain time on III, IV, V.
• The PM Oversight Angle pattern applies to Domain II too. Same oversight verbs, same documented-decision framing.
• II.8 (Success Criteria) is the most cross-pulled artifact in the exam. Master rigorous, measurable, business-tied criteria definition.
• Trustworthy-AI cost belongs in II.5 (ROI) and II.9 (business case). Don't omit.