Intro

But here’s where it gets kind of crazy.
Despite all that cash being thrown around, a whole bunch of studies estimate that a staggering 80% of AI projects, well, they failed to deliver what they promised.
80%.
Are we actually getting our money’s worth?

Is AI Actually Worth It? Why Most Companies Get AI ROI Wrong

Okay, so let’s be real.
It feels like everyone, and I mean everyone, is pouring money into AI right now.
But as those budgets get bigger and bigger, there’s this nagging question that’s getting a lot louder.
Are we actually getting our money’s worth?

Well, it turns out the way most companies measure the value of AI, it might be totally broken.
Let’s get into it.

AI 2028 spend prediction

First, you just have to appreciate the sheer scale of this thing.
The International Data Corporation predicts that by 2028, we’re going to be spending nearly a trillion dollars a year on AI globally.
That’s not just a trend, you know, that’s a full-blown title wave of investment.

But here’s where it gets kind of crazy.
Despite all that cash being thrown around, a whole bunch of studies, including a recent one from Rand, estimate that a staggering 80% of AI projects, well, they failed to deliver what they promised.
80%.

80% of AI projects fail

That is a massive amount of money going into projects that just don’t hit the mark.

So, what’s the deal?
Why is so much money leading to so much failure?

Well, here’s the twist.
It’s not always about bad technology.
A lot of the time, it’s about bad decisionmaking from the very beginning.

We’re picking the wrong projects because we’re measuring them with the wrong ruler.
And that brings us to what we can call the AI ROI paradox.

The AI ROI paradox

You see, the real heart of the issue is that we’re trying to measure this revolutionary complex technology using a super old school simplistic formula for return on investment.
It’s the classic square peg round hole problem.

Think about it like this.
That old ROI calculation, it’s like looking at just the tip of an iceberg.
It sees the obvious stuff, right?
The easy to count costs and savings.

But with AI, the real story, both the incredible potential and the pretty serious risks, it’s all hidden down below the surface.

What real AI value looks like: The real AI ROI

Okay, so let’s start with the return part of that ROI equation.
We have got to start looking beyond just simple cost carding because the true value that AI can bring to the table is so much bigger and more strategic than that.

So the traditional way of thinking is all about efficiency.
How many hours can we save?
How many dollars can we cut?

And that’s fine, but it’s just the beginning.

The real gamecher, the full picture is in its strategic value.
We’re talking about things that are a lot harder to slap a price tag on, like becoming more agile, making way smarter decisions, and building a competitive edge that honestly just leaves everyone else in the dust.

Now, check this out.
This quote just brilliantly captures a new way to think about value.
It introduces this concept of digital yield.

So, what it’s saying is, for every 4 hours an employee gets a little help from AI, the company gains what would have normally taken 3 weeks of work.

Just think about that for a second.
You’re not just saving a little time here and there.
You’re unlocking whole new levels of analysis and productivity that were basically impossible before.

That is a completely different kind of return.

The hidden AI costs no one budgets for

So now let’s flip over to the other side of that iceberg, the investment part.

Because the price you see on the box for an AI tool, that’s just the beginning.
The hidden costs are huge, and they’re exactly where project budgets go to die.

Let’s take a look under the water at these hidden costs.
You’ve got the enormous never-ending job of getting your data ready, cleaning it, labeling it, governing it.
Then you have these super complex usage based pricing models from vendors that are almost impossible to predict.

And you have to add in the costs for new security, for constantly tuning the model so they don’t drift.
And of course, the insane cost of finding and keeping people who actually know how to do all this.

These aren’t just one-time fees, you know.
They are deep recurring expenses that can absolutely sink a project.

But a real honest look at this can’t stop at hidden costs.
AI completely changes a company’s risk profile and we have to put a number on that.

This is about moving from a simple costbenefit thought process to a riskadjusted one.

A really powerful way to think about this is a concept called the risk delta.
It’s pretty simple when you break it down.

AI might reduce your risk of say financial fraud, which is great, but at the exact same time it might introduce brand new risks like a flawed algorithm causing some really expensive mistakes.

The risk delta is just the net financial impact of all those changes.
It forces you to ask that tough question.
Are we just trading one big problem for another?

And believe me, these new risks have very real teeth.

AI new risks

Take the EU’s AI Act for example.
If you don’t comply, you could be looking at penalties up to 35 million or even 7% of your global turnover.

All of a sudden, the cost of a new AI risk isn’t some abstract idea.
It’s a massive potential liability that has to be part of the math.

The AI investment playbook

So, knowing all this, how on earth can anyone make better decisions?

Well, it takes a whole new playbook.
One that’s more strategic, more holistic, and constantly evolving.

Okay, so here’s what that modern playbook looks like.

First, you’ve got to adopt a truly holistic analysis.
That means you toss out that simple ROI formula and you start including strategic benefits and the entire cost of ownership.

Second, you start treating your AI projects like a balanced investment portfolio.
Some are safe bets for quick wins and others are those big moonshots aiming for major transformation.

Then you evaluate constantly.
This isn’t a one-time thing.
You have to measure everything from how happy your users are to the cost of every single interaction at every stage.

And that leads to the final and maybe most important step.
You scale intelligently.

You set aside budgets just for experimenting.
You start small.
Prove the value with real data.
And then and only then do you grow based on what you’ve actually learned, not just on the initial hype.

The real question leaders should ask about AI ROI

And that really brings us to the final thought here.

The goal shouldn’t just be to calculate a simple return on investment.
The real question you should be asking is, are you building a lasting strategic advantage?

Are you ready to move beyond basic arithmetic and start measuring your company’s intelligence dividend?

Because in the age of AI, that’s the number that’s really going to define who wins.