Strategic Ai Investment

AI-Driven Business Innovation Masterclass

Strategic AI Investment: A Portfolio Approach

Executive Summary

Rather than asking “Should we invest in AI?” (the answer is almost certainly yes), executives must ask: “Where should we invest in AI, and how should we balance our portfolio?” Use McKinsey’s Three Horizons model: allocate 70% to Horizon 1 (optimise existing operations, 0-12 months, proven tech), 20% to Horizon 2 (build emerging capabilities, 1-3 years, scaling challenges), and 10% to Horizon 3 (transform business models, 3-5+ years, high risk). Evaluate AI investments across four value categories—cost reduction, revenue growth, risk reduction, and strategic positioning—not just traditional ROI. Over-investing in H1 optimises you into obsolescence; over-investing in H3 starves near-term revenue needed to fund experiments.

Reading time: 15 minutes Key takeaway: AI portfolio balance matters more than individual project ROI—use the 70/20/10 rule across three time horizons.

The Portfolio Mindset

Traditional project evaluation asks: “Is this project worth funding?”

AI portfolio management asks: - How does this fit with our other AI investments? - What’s the right balance across time horizons? - Where are we taking concentrated risk? - What capabilities are we building for the future?

Why Portfolio Thinking Matters

The problem with project-by-project evaluation: - Creates gaps in capability building - Over-invests in safe, incremental improvements - Under-invests in transformational opportunities - Misses competitive threats from adjacent markets

The portfolio approach: - Balances risk across initiatives - Builds complementary capabilities - Maintains optionality for future moves - Creates strategic coherence

Three Time Horizons

McKinsey’s Three Horizons model provides a framework for balancing AI investments:

Horizon 1: Optimise Core (0-12 months)

Focus: Improve existing operations
Budget allocation: 70%
ROI expectation: >200% in Year 1
Risk profile: Low risk, proven technology

Examples: - Process automation (RPA + AI) - Predictive maintenance - Customer service chatbots - Demand forecasting

Questions to ask: - Does this make our core business more efficient? - Can we implement this with existing capabilities? - Will this free up resources for higher-value work?

Horizon 2: Build Emerging (1-3 years)

Focus: Grow adjacent opportunities
Budget allocation: 20%
ROI expectation: >150% over 3 years
Risk profile: Medium risk, scaling challenges

Examples: - Personalisation engines - Dynamic pricing optimisation - Supply chain AI - Fraud detection

Questions to ask: - Does this create new revenue streams? - Are we building capabilities competitors don’t have? - Can we scale this if successful?

Horizon 3: Create Future (3-5+ years)

Focus: Transform business model
Budget allocation: 10%
ROI expectation: >100% over 5 years
Risk profile: High risk, high potential

Examples: - AI-powered products - Autonomous operations - Platform business models - Industry disruption plays

Questions to ask: - Could this redefine our industry? - Are we building options for the future? - What happens if we don’t invest and competitors do?

Four Value Categories

AI creates value in four ways. Strong business cases articulate value across multiple categories:

1. Cost Reduction

Direct savings: Labour, materials, overhead
Efficiency gains: Time savings, waste reduction
Easily quantifiable, but often oversold

Warning: Don’t evaluate transformational AI purely on cost savings. That’s like evaluating the internet based on saving postage costs.

2. Revenue Growth

New offerings: AI-powered products/services
Market expansion: Reach new customers
Pricing power: Charge premium for AI features

Challenge: Harder to quantify, but often the bigger prize.

3. Risk Reduction

Fraud prevention: Detect anomalies
Compliance: Automated monitoring
Quality control: Reduce defects/errors

Often overlooked: Risk reduction is hard to measure but can be existential.

4. Strategic Positioning

Competitive advantage: Capabilities competitors can’t match
Market leadership: First-mover advantages
Talent attraction: Best people want to work with cutting-edge tech

Impossible to quantify: But may be the most important long-term value.

The Investment Decision Framework

Evaluation Criteria

For Horizon 1 (Optimise): 1. Clear ROI >200% in Year 1 2. Low technical risk 3. Fast implementation (<6 months) 4. Builds foundational capabilities

For Horizon 2 (Emerge): 1. Substantial revenue potential 2. Defensible competitive advantage 3. Scalability path is clear 4. Manageable implementation risk

For Horizon 3 (Transform): 1. Could redefine industry dynamics 2. Builds critical future capabilities 3. Creates strategic options 4. Risk is acceptable given portfolio balance

Red Flags

Warning signs in AI proposals: - “AI will solve everything” (too vague) - No clear success metrics - Assumes perfect data quality - Ignores organisational change requirements - “Everyone else is doing it” (FOMO-driven) - Over-confidence in ROI projections - Underestimates integration complexity

Green Lights

Strong AI investment proposals: - Clear problem definition - Specific, measurable success criteria - Realistic data assessment - Change management plan - Pilot approach with clear scale criteria - Multiple value categories - Fits portfolio strategy

Common Executive Mistakes

Mistake 1: Treating AI Like Traditional IT

Problem: AI requires experimentation, not waterfall planning
Solution: Fund pilots, not full implementations. Build learning cycles.

Mistake 2: Underinvesting in Data Infrastructure

Problem: Fancy algorithms fail without quality data
Solution: 30-40% of AI budget should go to data infrastructure.

Mistake 3: All Horizon 1, No Horizon 3

Problem: Optimise into obsolescence while competitors transform
Solution: Force yourself to allocate 10% to moonshots.

Mistake 4: Overestimating Short-term Impact

Problem: AI takes longer to implement than vendors promise
Solution: Double the timeline, halve the initial ROI projection.

Mistake 5: Underestimating Long-term Impact

Problem: Strategic positioning value is often undervalued
Solution: Consider what your industry looks like if you don’t invest.

Competitive Dynamics

The AI Arms Race

First-mover advantages: - Data accumulation (the flywheel effect) - Talent acquisition (best people join leaders) - Customer lock-in (AI improves with usage) - Brand perception (innovation leader)

Fast-follower advantages: - Learn from pioneers’ mistakes - Lower technology risk - Better ROI clarity - Avoid dead-end investments

Know which game you’re playing: Are you defining a new category (first-mover) or optimising a proven approach (fast-follower)?

Disruptive Threats

Watch for: - Startups with AI-first business models - Big tech entering your industry - Adjacent industry players with AI capabilities - Changing customer expectations driven by AI

The question isn’t “Will AI disrupt our industry?”
The question is “Will we be the disruptor or the disrupted?”