Executives Should Start AI Investments With Strategy, Not Tools
Leaders waste money on AI pilots that solve no real problem. The fix: map concrete business bottlenecks first, then decide if AI fits.
AI strategist Justin Massa outlines a framework for executives to prioritize AI spending where it delivers measurable returns. The approach starts by identifying high-frequency, high-cost processes that create friction-then designing narrow pilots around those specific pain points.
Most organizations reverse this order. They buy tools, then hunt for uses. That approach generates abandoned projects and inflated tooling budgets.
The AI Opportunity Matrix
Massa proposes scoring initiatives across four categories: Maintain, Automate, Augment, and Transform. Each sits on a grid measuring business importance against exposure to AI disruption.
- Maintain: Keep current processes as-is. AI adds no advantage.
- Automate: Replace manual work with AI. High efficiency gain, low risk.
- Augment: Pair AI with human judgment. Improves speed or quality of existing work.
- Transform: Reimagine a process entirely. Highest reward and highest disruption risk.
The Transform quadrant demands careful attention. These bets carry the most uncertainty and require deliberate governance.
How to Run Pilots That Matter
Practitioners should follow four concrete steps:
- Map processes by frequency and cost. Quantify current throughput or error rates so you can measure improvement.
- Design pilots with narrow scope. Solve one bottleneck, not five. Define success metrics before you start.
- Assemble cross-functional teams. Combine domain experts, product managers, and engineers for rapid iteration.
- Establish measurement guardrails. Set quality thresholds and human-in-the-loop checkpoints for model outputs.
This prevents the common trap: a six-month pilot that proves nothing because success was never defined.
Generative AI Works Today-Don't Wait for Tomorrow
"When I think about what you should be doing with generative AI right now, it's really about leveraging what the technology can do today, not what it can do tomorrow," Massa said.
Organizations that chase speculative capabilities delay action. The practical move is deploying current generative AI strengths-content generation, summarization, code assistance, customer support triage-against real problems now.
This shifts the conversation from hype to outcomes. Does the pilot reduce time spent on a task? Does it cut error rates? Does it improve customer experience? Tie every investment to one of those metrics.
Balance Efficiency With Creative Work
Not all AI use cases are about automation. Augmentation-pairing AI with human creativity-often delivers higher business value than pure efficiency plays.
The framework works because it mirrors established strategy practices. Organizations already familiar with Roger Martin's Playing to Win approach will recognize the structure. That familiarity makes it easier to integrate AI decisions into existing planning cycles, rather than treating AI as a separate strategic exercise.
The real measure of success is whether teams convert pilots into repeatable patterns and operationalized platforms for model lifecycle management. One-off experiments prove nothing. Scaled, measurable adoption is what separates winners from the rest.
Learn more about AI for Executives & Strategy and Generative AI and LLM capabilities.
Your membership also unlocks:
