Stop Measuring AI on Savings, Start Measuring Its Strategic Value.

For some, AI is only about statistics of failure.

The others have the plan, implement AI, and they are winning.

What's the difference?

They stopped counting pennies saved and started building the future.

Before Measuring the Future, Imagine It

You don't have time to think strategically. You ask for costs and ROI like every other project. You isolate AI use cases without considering the broader context.

But innovation doesn't work that way.

Close your eyes and imagine your company in the future utilizing AI to its maximum potential. I mean it. Just five minutes. Now imagine spending a week imagining the future with your entire management board.

That's where real strategy begins.

The companies that win don't start with spreadsheets. They start with a clear room for thinking. Five to seven days. Off-grid. No laptops during daylight. Phones in a Faraday pouch. Full management board plus domain experts. One clear objective: reimagine your business with AI, then come back with a shared future, not a vague intention.

This requires scanning the external environment and your organization's internal landscape, then generating vivid pictures of possible futures. You need to understand what becomes instant, what becomes infinite, and where autonomy shows up first in your specific context.

Most organizations stall not because the tech is hard, but because leadership alignment is missing. When executives don't set direction and frame trade-offs, teams collect tools, not outcomes.

The fix isn't more demos. It's a coherent blueprint that connects market drivers to ambition, roadmap, and enablers. When you explicitly tie outside-in market drivers to future-back intent, then to principles that guide your execution, pilots roll up to strategy instead of dying in isolation.

Design Smart KPIs to Win in the Long Term

Imagine you decided to start running every day. The simplest KPI might be how much faster you will run over one kilometer after 30 days.

But is that all?

No.

There are indirect, positive side effects:

- Your pulse will lower and your blood pressure will decrease, so your heart health improves
- You'll probably sleep better because running expends energy and reduces cortisol
- Your skin will look better because running increases circulation to the brain and nourishes cells with oxygen
- Your cognitive function improves, including executive function and decision-making abilities
- Your mood and mental health improve through the release of endorphins and serotonin
- Your stress response resets, training your nervous system to bounce back more quickly
- Your self-esteem and confidence boost with each milestone achieved

This is what I mean by smart KPIs.

The Real AI Use Case Example

Let's imagine a real scenario with a single AI use case. You want to automate the manual data entry from handwritten timesheets of temporary workers into a database, hundreds of thousands per year. Let's imagine this solution is already implemented.

Instead of manual, you have a partially automated process. Instead of 50 employees entering data, you need only 40.

That's your direct KPI—productivity increase.

But what are the indirect benefits of this implementation?

1. AI Maturity

If it's your first initiative, you've created for the first time an AI environment for process automation. Maybe a cloud environment. Perhaps a relation with an AI vendor. You have automation tools and platforms in place. This is foundational infrastructure that compounds across future projects.

2. Knowledge Management

For the first time, you documented the process you're performing in the organization. You gain a deeper understanding of how your organization operates behind the scenes. You're more resilient. If employees leave, your knowledge stays.

3. Data Quality

This project influenced your data collection. It's now better organized. Instead of receiving documents through five different channels—personal email, general email, random folders—it simplifies through one AI document processing pipeline. Your data infrastructure just leveled up.

4. Process Quality

Because of these indirect KPIs, you improved the process itself. It's better organized. It's documented. Even after removing AI, it should work better. You increased organizational order.

5. Innovation Culture

You increased awareness of AI technology across the IT department. Not only maturity, but also openness to new areas of innovation. Your technical teams now see what's possible. They start proposing ideas.

6. Change Management

You increased awareness of AI among regular employees. You can start managing this dynamic. Some will be afraid. Some will embrace it. You can start managing the change and seeing the benefits and threats from the psychological point of view.

How to Assign Tangible Value

Now the hard part: how to assign tangible value to each of these indirect KPIs.

Here's my proposal:

Total Direct Benefit: $300K annually (10 FTEs saved at $50K each, assuming use case cost $200K)

Total Indirect Benefit: $3M over 3 years

The ratio speaks for itself. 

Is it harder to measure? Yes.

It's harder to be a winner. Choose your hard.

Select the Proper Measure for Each of the AI Waves

Now, let's move beyond a single use case. Let's imagine the future development of process automation across three AI adoption waves.

Wave 1 focuses on time, cost, and efficiency. You apply AI to existing workflows to do what you already do, but faster and cheaper. Wave 2 elevates quality. AI doesn't just speed things up, it makes things better. Wave 3 transforms the system entirely. This isn't about improving your current business model—it's about making it obsolete before someone else does.

Let's focus on automating one manual process through all three waves.

Notice the pattern.

Wave 1 delivers immediate ROI but a temporary competitive advantage. Every competitor can do this. Most already are.

Wave 2 builds strategic capabilities that compound. The best companies are at this frontier now with AI agent experiments, augmenting human expertise, and elevating quality.

Wave 3 creates the moat. You need to be doing all three simultaneously. Success in Wave 1 creates the conditions for Wave 3 failure if you're not careful. You optimize existing processes, get better at what you already do, and your culture reinforces the current business model. Then someone builds a Wave 3 system that makes your entire operation obsolete.

The Template You Can Steal

Here's a simple framework to map any AI initiative across waves with both direct and smart KPIs.

Super simple. Steal it and customize it.

Here is the link.

The Bottom Line

Financial justifications for initiatives have become and will continue to become more important, particularly for larger, more expensive, and more strategically important ones.

But if you're only measuring cost savings, you're optimizing for yesterday.

By 2030, AI won't just support your business. It will be your business. The winners won't be the fastest adopters, but the boldest architects.

Stop measuring AI on what it saves.

Start measuring what it builds.

The companies winning today didn't just deploy AI faster. They imagined a different future, designed smart KPIs across all three waves, and built the infrastructure to compound strategic value over time.

They understood that the ratio between direct and indirect benefits is often 1:3 or more. They knew that Wave 1 efficiency pays for Wave 2 quality, which enables Wave 3 transformation.

Most importantly, they stopped asking their CFO for ROI justification on individual use cases and started showing their board a coherent strategy for becoming an AI-native business.

That's the difference between statistics of failure and those of success.

Your move.