Chapter 10 — Business Case & Sponsorship

Overview

Quantify value, costs, and risks to secure executive sponsorship and governance. The business case is the foundation for AI investment decisions—it translates technical capabilities into business outcomes and provides sponsors with the clarity they need to commit resources.

A strong business case enables fast decisions, establishes accountability, and sets the stage for successful AI delivery. This chapter shows how to build compelling, evidence-based business cases that secure buy-in and minimize surprises.

The Three-Part Business Case Model

Every AI business case must address value, cost, and risk with equal rigor:

graph TB
    A[Business Case] --> B[Value Model]
    A --> C[Cost Model]
    A --> D[Risk Model]

    B --> B1[Revenue Impact]
    B --> B2[Cost Savings]
    B --> B3[Risk Reduction]

    C --> C1[Build Costs]
    C --> C2[Run Costs]
    C --> C3[Change Costs]

    D --> D1[Delivery Risk]
    D --> D2[Model Risk]
    D --> D3[Compliance Risk]

    B1 --> E[NPV Calculation]
    B2 --> E
    B3 --> E
    C1 --> E
    C2 --> E
    C3 --> E

    E --> F[ROI Metrics]
    D1 --> F
    D2 --> F
    D3 --> F

    F --> G[Go/No-Go Decision]

    style B fill:#d4edda
    style C fill:#fff3cd
    style D fill:#ffe1e1

Component 1: Value Model

Value Streams

Revenue Generation:

Cost Savings:

Risk Reduction:

Building the Value Model

Step 1: Establish Baselines

## Baseline Data Collection

### Metric: Conversion Rate
- **Source**: Google Analytics, 12-month history
- **Current Performance**: 2.5% (range: 2.3-2.7% seasonal)
- **Data Quality**: High (100K+ samples/month)
- **Stability**: Stable, no major trends

### Metric: Customer Service Volume
- **Source**: Zendesk, 24-month history
- **Current Performance**: 2,500 tickets/month, 45-min avg handle time
- **Data Quality**: High (complete records)
- **Stability**: Growing 5%/year, seasonal Q4 peaks

Step 2: Estimate Impact with Ranges

Use three scenarios: Conservative, Base, Optimistic

Probability Weighting: (0.25 × $450K) + (0.50 ×$675K) + (0.25 × $945K) = **$686K expected value**

Step 3: Model Adoption Curve

graph LR
    A[Month 0-3<br/>Alpha: 5% users<br/>$5K/month value] --> B[Month 4-6<br/>Beta: 25% users<br/>$25K/month value]
    B --> C[Month 7-9<br/>Rollout: 60% users<br/>$60K/month value]
    C --> D[Month 10-12<br/>Steady: 85% users<br/>$85K/month value]
    D --> E[Year 2+<br/>Optimize: 90% users<br/>$90K/month value]

    style E fill:#d4edda

Cumulative Value: Months 1-3: $15K + Months 4-6:$75K + Months 7-9: $180K + Months 10-12:$255K = **$525K Year 1** (vs.$1.02M if instant 85% adoption)

Step 4: Sensitivity Analysis

Insight: Focus de-risking on adoption (training, change management) and validating conversion lift (A/B tests).

Component 2: Cost Model

Total Cost of Ownership (TCO)

Build Costs:

Run Costs (Annual):

Build vs. Buy vs. Partner Analysis

graph TD
    A[Solution Decision] --> B{Core Competency?}
    B -->|Yes| C{Have Capability?}
    B -->|No| D[Buy or Partner]

    C -->|Yes| E[Build]
    C -->|No| F{Time to Acquire?}

    F -->|<6 months| G[Build + Hire]
    F -->|>6 months| D

    D --> H{Differentiation?}
    H -->|High| I[Partner - maintain control]
    H -->|Low| J[Buy - standardize]

    E --> K[Pros: Control, IP, customization<br/>Cons: Time, cost, risk]
    G --> L[Pros: Capability building<br/>Cons: Hiring risk, time]
    I --> M[Pros: Speed, expertise, shared risk<br/>Cons: Dependency, cost]
    J --> N[Pros: Fast, proven, support<br/>Cons: Less differentiation, ongoing cost]

Decision Matrix:

Recommendation: Partner approach scores highest—balance of speed, cost, and control.

Component 3: Risk Model

Risk Taxonomy

Delivery Risks:

Model Risks:

Compliance Risks:

Risk-Adjusted Financial Model

Scenario Planning:

Interpretation: Even with 35% chance of underperformance or failure, expected NPV is positive ($905K). However, 10% chance of total loss means strong governance and kill criteria are essential.

Sponsorship & Governance

Sponsor Role Definition

Executive Sponsor Charter:

## Executive Sponsor Charter

### Role Purpose
Champion the AI initiative, remove organizational blockers, ensure strategic alignment.

### Decision Rights (RACI: Accountable)
- Approve business case and budget allocation
- Resolve cross-functional conflicts and escalations
- Approve go-live and major scope changes
- Final say on kill/pivot decisions at phase gates

### Time Commitment
- 2-4 hours/month during discovery/build
- 30 minutes/week during critical phases
- Monthly steering committee (1 hour)
- Available for urgent escalations (<24hr response)

### Success Criteria
Project delivers [specific value targets] within [timeline] and [budget],
with [quality metrics] met.

### Sponsor: [Name, Title]
### Term: [Start Date] to [End Date or "until completion"]

Governance Structure

graph TB
    A[Steering Committee<br/>Monthly, 1hr<br/>Sponsor + C-suite] -->|Strategic guidance<br/>Budget approval| B[Core Team<br/>Weekly, 1hr<br/>PM + Tech + Product]

    B -->|Progress updates<br/>Escalations| A

    B -->|Coordination| C[Data Workstream<br/>2x/week, 30min]
    B -->|Coordination| D[Engineering Workstream<br/>Daily standup, 15min]
    B -->|Coordination| E[Change Mgmt Workstream<br/>Weekly, 30min]

    C -->|Data readiness<br/>Quality reports| B
    D -->|Dev progress<br/>Technical risks| B
    E -->|Adoption metrics<br/>User feedback| B

    F[RAID Log<br/>Tracked in Jira] -->|Risks| B
    F -->|Actions| C
    F -->|Actions| D
    F -->|Actions| E

    G[Phase Gates<br/>Quarterly] -->|Go/No-Go| A
    B -->|Gate materials| G

    style A fill:#e1f5ff
    style B fill:#fff3cd
    style G fill:#d4edda

Steering Committee Operating Model:

Deliverables

1. Business Case Document

Executive Summary (1 page):

• Opportunity: [What problem are we solving?]
• Solution: [How will AI address it?]
• Value: [Expected benefits with ranges]
• Investment: [Total cost over 3 years]
• ROI: [Payback period, NPV, IRR]
• Risks: [Top 3 risks and mitigations]
• Recommendation: [Go/No-Go with rationale]

Value Model (2 pages):

Cost Model (2 pages):

Risk Model (2 pages):

• Top 10 risks with probability, impact, mitigation, owner
• Scenario analysis (Best/Base/Worse/Failure with probabilities)
• Contingency plan and kill criteria

2. Sponsor Brief (1-Pager)

# Sponsor Decision Brief: [Initiative Name]

**Date**: [Date]
**Decision needed by**: [Date]

## The Ask
[Clear, specific request: e.g., "Approve $540K to build AI customer service assistant"]

## The Business Case
| Metric | Value |
|--------|-------|
| **Investment** | $540K build + $460K/year run = $1.86M over 3 years |
| **Return** | $2.95M value over 3 years |
| **NPV** | $905K (3-year, 12% discount rate) |
| **Payback** | 9 months after launch |
| **Confidence** | Medium (50% probability of base case) |

## Key Risks & Mitigations
1. **Model accuracy**: 35% probability → Baseline testing, fallback to rules, human-in-loop
2. **Adoption slower**: 40% probability → Change management, phased rollout, champions
3. **Integration challenges**: 40% probability → POC completed successfully, dedicated engineer

## Phase Gates (Kill/Pivot Criteria)
- **Gate 1 (Month 3)**: Data quality >85%, model accuracy >target
- **Gate 2 (Month 6)**: Integration complete, UAT passed
- **Gate 3 (Month 9)**: Pilot shows >60% adoption, positive NPS

## Recommendation
**GO** - Proceed with build phase
- Strong business case with positive NPV in risk-adjusted scenarios
- Key technical risks mitigated through POCs
- Competitive pressure makes this strategically important

**Sponsor Decision**: [ ] Approve  [ ] Modify:  [ ] Reject

**Signature**: _________________ **Date**: _______

Case Study: AI Customer Service Assistant

Context: Mid-sized insurance company wants AI assistant to handle routine inquiries, reducing call volume and improving response times.

Value Model

Baseline Data:

• Call volume: 50,000 calls/month
• Average handle time: 12 minutes
• Cost per call: $8 (agent loaded cost)
• Call center budget: $4.8M/year
• Current CSAT: 3.2/5

Value Calculation (Base Case):

Adoption Curve:

• Months 1-3: 10% (beta) = $225K
• Months 4-6: 40% (rollout) = $900K
• Months 7-12: 70% (steady state) = $1.58M
• Year 1 realized value: $2.7M cumulative

Cost Model

• Build: $540K (platform setup$60K, integration $120K, training data$80K, model tuning $100K, change mgmt$90K, contingency $90K)
• Annual Run: $304K (LLM API$109K, infrastructure $45K, support$90K, monitoring $20K, contingency$40K)
• 3-Year TCO: $1.41M

Financial Summary

Scenario Analysis:

Outcome

Sponsor Decision: APPROVED with conditions

• Proceed to build phase
• Gate 1 criteria must be met before full budget commit
• Weekly status during critical months 4-9
• Hard stop if CSAT drops >0.2 points

Results (12 months post-launch):

• Actual call deflection: 28% (vs. 30% target) - close enough ✅
• CSAT improved 0.4 points (exceeded 0.3 target) ✅
• Agents initially resistant but became advocates after easier workload ✅
• Payback in 10 months (vs. 9-month projection) ✅
• Year 1 value: $2.4M (vs.$2.25M projected) ✅

Why It Worked:

1. Conservative assumptions: Didn't overpromise, built credibility
2. Evidence-based: POC validated accuracy before full commit
3. Risk-aware: Identified and mitigated key risks (agent resistance, accuracy)
4. Strong sponsor: SVP provided air cover during resistance periods
5. Clear governance: Monthly steering kept project on track, caught issues early

Implementation Checklist

Phase 1: Preparation (Weeks 1-2)

• Identify value drivers through stakeholder interviews
• Gather baseline data for all proposed metrics
• Research benchmarks and analogies for estimates
• Document current costs as comparison baseline

Phase 2: Value Model (Weeks 3-4)

• Build value model with conservative/base/optimistic scenarios
• Validate assumptions with domain experts
• Model adoption curve based on similar initiatives
• Run sensitivity analysis on key assumptions
• Determine if pilot or POC is worthwhile (VOI analysis)

Phase 3: Cost Model (Weeks 4-5)

• Enumerate all cost categories (build, run, change)
• Get vendor quotes for platforms and services
• Estimate team needs and loaded costs
• Build vs. buy vs. partner analysis
• Add contingency buffers (15-20%)

Phase 4: Risk Model (Week 5)

• Identify delivery, model, and compliance risks
• Assess probability and impact for each risk
• Develop mitigation plans with owners and costs
• Create scenario analysis (best/base/worse/failure)
• Calculate risk-adjusted NPV

Phase 5: ROI Synthesis (Week 6)

• Calculate NPV, IRR, payback period
• Build decision tree or tornado chart
• Validate calculations with finance team
• Pressure-test assumptions with skeptics

Phase 6: Sponsorship (Week 7)

• Identify and recruit executive sponsor
• Draft sponsor charter with clear decision rights
• Secure sponsor commitment (time and authority)
• Identify product owner and other key roles

Phase 7: Governance Design (Week 7)

• Design governance structure (steering, core team)
• Define RACI for key decisions
• Establish meeting cadence and agendas
• Set up RAID tracking system
• Define phase gate criteria and kill triggers

Phase 8: Documentation (Week 8)

• Draft full business case document
• Create 1-page sponsor brief
• Write governance charter
• Prepare presentation deck
• Assemble appendices (detailed models, research)

Phase 9: Socialization (Weeks 8-9)

• Pre-brief sponsor 1:1 before formal review
• Incorporate sponsor feedback
• Present to steering committee
• Address questions and concerns
• Revise based on feedback

Phase 10: Decision & Launch (Week 10)

• Obtain formal approval and signatures
• Communicate decision to broader team
• Kick off governance (schedule first steering)
• Set up tracking and reporting cadence
• Begin execution with clear success criteria

Ongoing: Track & Update

• Update financial model quarterly as actuals come in
• Revise risk register as risks materialize
• Report value realization vs. plan monthly
• Adjust forecast based on learnings
• Conduct post-launch review to validate assumptions

---

Key Takeaway: A strong business case is evidence-based, risk-aware, and decision-ready. It quantifies value with ranges not false precision, acknowledges costs holistically including change management, and treats risks explicitly with mitigation plans. Combined with clear sponsorship and governance, it sets the foundation for AI initiatives to deliver on their promise.