AI Use Case Assessment Worksheet

Use Case Name: _________________________________

Date: _________________________________

Assessor: _________________________________

Dimension 1: Desirability — Is the Problem Worth Solving?

Problem Statement (Without Mentioning AI)

What is the actual business problem?

Who owns this problem?

What's the current cost of this problem?

Time cost: _________________________________
Money cost: _________________________________
Error cost: _________________________________

What's the cost of doing nothing?

What does "solved" look like?

Success Metrics

Baseline Performance (Current State):

Metric 1: _________________________________ = _________________________________
Metric 2: _________________________________ = _________________________________
Metric 3: _________________________________ = _________________________________

Target Performance (Success State):

Metric 1: _________________________________ = _________________________________
Metric 2: _________________________________ = _________________________________
Metric 3: _________________________________ = _________________________________

How will we measure success?

Quantitative metric (specify): _________________________________
Qualitative feedback (specify): _________________________________
Business impact (specify): _________________________________

Strategic Alignment

Aligns with business priorities
Executive sponsorship confirmed
Users are engaged and supportive
Clear business owner identified

Desirability Score (1-10)

Criterion	Score (1-10)	Notes
Quantified Impact	___ / 10
Strategic Alignment	___ / 10
Success Metrics Clarity	___ / 10
Business Owner Commitment	___ / 10
TOTAL	___ / 40

Desirability Decision:

✅ PASS (Score ≥30): Problem is worth solving
🟡 REVIEW (Score 20-29): Needs refinement
❌ FAIL (Score <20): Not worth solving

Dimension 2: Feasibility — Can We Technically Do This?

Data Reality Check

Do we HAVE the data we need? (Not "can we collect it"—do we HAVE it?)

Yes, data exists
No, data doesn't exist
Partial, some data exists

Data Details:

Data source: _________________________________
Data owner: _________________________________
Historical depth: _________________________________
Volume: _________________________________ records
Update frequency: _________________________________

Data Quality:

Labeled training data available
Data completeness: _____% (target: >85%)
Data accuracy: _____% (target: >85%)
Data freshness: _____ (how recent?)
Missing values: _____% (target: <15%)

Data Access:

Legal/compliance approval obtained
Privacy review completed
Data owner has granted access
Technical access is feasible

Technical Fit

Does this problem require AI, or would rules/heuristics work?

Problem is ambiguous/probabilistic → AI needed
Rules can fully capture logic → No AI needed
Some error is acceptable → AI suitable
Pattern exists in historical data → AI suitable
Humans struggle to scale decisions → AI suitable

Simpler Solutions Tested:

_________________________________ → Result: _________________________________
_________________________________ → Result: _________________________________
_________________________________ → Result: _________________________________

Why simpler solutions are insufficient:

Technical Skills:

Team has required ML/AI skills
Can acquire skills through training
Need to hire external expertise
Skills gap is a blocker

Integration Feasibility:

Can integrate with existing systems
No major technical blockers
Infrastructure is ready
API/access available

Compliance & Governance

GDPR/compliance review completed
Security sign-off obtained
Data residency requirements met
Industry regulations considered
Audit logging requirements defined

Feasibility Score (1-10)

Criterion	Score (1-10)	Notes
Data Availability	___ / 10
Data Quality	___ / 10
Technical Skills	___ / 10
Integration Feasibility	___ / 10
Compliance Readiness	___ / 10
TOTAL	___ / 50

Feasibility Decision:

✅ PASS (Score ≥40): Technically feasible
🟡 REVIEW (Score 30-39): Needs PoC to validate
❌ FAIL (Score <30): Not feasible right now

Dimension 3: Viability — Can We Sustain This?

ROI Calculation

Annual Benefit (What will be saved/earned):

Benefit 1: _________________________________ = €_____ / year
Benefit 2: _________________________________ = €_____ / year
Benefit 3: _________________________________ = €_____ / year
Total Annual Benefit: €_____ / year

Implementation Cost (One-time):

Development: €_____
Infrastructure setup: €_____
Team training: €_____
Total Implementation Cost: €_____

Operating Cost (Ongoing, per year):

Infrastructure: €_____ / year
Team (FTE): €_____ / year
Maintenance: €_____ / year
Total Operating Cost: €_____ / year

Payback Period:

Year 1: €_____ (benefit) - €_____ (implementation) - €_____ (ops) = €_____
Year 2: €_____ (benefit) - €_____ (ops) = €_____
Payback achieved in: _____ months

Risk-Adjusted Scenario (50% as good):

Annual benefit (50%): €_____ / year
Payback period: _____ months
Still positive? [ ] Yes [ ] No

Team & Skills

Team is assigned
Team has required skills
Training plan is defined
Can maintain long-term
Succession plan exists

Change Management

Users are engaged
Process changes are defined
Training plan for users
Adoption strategy exists
Organization is ready

Viability Score (1-10)

Criterion	Score (1-10)	Notes
ROI Positive	___ / 10
Payback Period	___ / 10
Team Readiness	___ / 10
Change Management	___ / 10
TOTAL	___ / 40

Viability Decision:

✅ PASS (Score ≥30): Financially justified
🟡 REVIEW (Score 20-29): Needs refinement
❌ FAIL (Score <20): Not financially justified

AI Level Selection

Recommended AI Level:

Level 0: No AI needed (rules/heuristics)
Level 1: Analytics/BI (statistical models)
Level 2: AI-Supported (human decides, AI suggests)
Level 3: AI-Integrated (AI decides, embedded in workflow)
Level 4: Advanced ML (real-time, complex patterns)
Level 5: Agentic AI (autonomous execution)

Why this level?

Estimated Cost: €_____

Estimated Timeline: _____ weeks/months

Can we start simpler and upgrade later?

Yes, start at Level _____
No, need this level from day one

Overall Assessment

Summary Scores

Dimension	Score	Decision
Desirability	___ / 40	✅ / 🟡 / ❌
Feasibility	___ / 50	✅ / 🟡 / ❌
Viability	___ / 40	✅ / 🟡 / ❌
TOTAL	___ / 130