Why Business Value is the Wrong Metric for AI Projects

Picture this: You're a commercial real estate developer. Everyone advises you to pick the project with the highest rental or purchase yields. Sounds logical, right? But, when you dig deeper.

• What if your foundation can't support the structure?
• What if you don't have the specialized crews needed?
• What if your building materials aren't readily available?

This is exactly what's happening in enterprise AI right now.

Every Friday morning, I deliver 1 actionable insight to help you navigate the post-AI landscape, simplifying complex transitions into a clear path for business impact. You can click here to subscribe.

I've spent over a decade deploying AI applications across Fortune 500 companies, and I keep seeing the same mistake: organizations choosing AI projects based purely on business value.

It's like choosing to build a luxury high-rise without checking if you have the right foundation, expertise, or materials – it looks great in the proposal, but often crumbles in reality.

I was brought in to assess a company's AI strategy. They had 12 use cases, all with compelling business value calculations. The management consulting firm that developed them had even put million-dollar figures next to each one. It was so irresistible, but sadly, misleading.?

After digging deeper, I shared the uncomfortable truth: 9 of those 12 projects were actually impossible to implement with their current resources.?

Business value is an alluring but dangerous metric when used alone.

After years of face-planting (yes, I've got the scraped knees to prove it), I developed what I call the Criticality-Complexity Matrix.

This framework is copyrighted: ? 2025 Sol Rashidi. All rights reserved.

Here's how it works:

Criticality Questions To Ask (this is a subset and not a complete list):

• Is there an imminent threat from competitors?
• Are regulations and fines looming?
• Are you losing market share?
• Can you quantify the immediate impact?

Complexity Questions To Ask (this is a subset and not a complete list):

• Do you have the basic infrastructure?
• Is your data accessible and usable?
• Do you have the right talent available?
• Can your organization adapt operationally?

Then each question gets a weight, and that weight is given based on the answer given - and through a simple formula it gets plotted on a quadrant that lets you visually see which use cases are:

• No Brainers (can easily be executed)
• Avoid (cannot be executed)
• Requires Planning (can be executed but it’s complicated and requires time and investments)
• Optional (only pursue but for exceptional reasons)

The full walkthrough of the Criticality-Complexity Matrix is in my bestselling book - Your AI Survival Guide. Check it out, and I’m happy to chat if you’re looking for some help with your AI strategy.?

The results often surprise executives – what looked like a high-value project suddenly reveals itself as a high-complexity nightmare waiting to happen

Achievement creates momentum.

Success in AI isn't about chasing the biggest numbers – it's about understanding what your organization can realistically execute and start building from there.

Just like in real estate development, you don't break ground on a luxury high-rise for your first project. You start with manageable developments, perfect your construction processes, and build a reliable team. Then, as your capabilities grow, you can tackle those more ambitious towers on the horizon.

Let me know if this resonates with your experience. :) What metrics do you use to evaluate AI projects?

Every Friday morning, I deliver 1 actionable insight to help you navigate the post-AI landscape, simplifying complex transitions into a clear path for business impact. You can click here to subscribe.