The ROI of AI: When Should It Show Up in EBITDA?

Almost every CEO I speak with is investing in AI.

Some are deploying it across finance. Others are implementing it in sales, marketing, operations, or customer support.

The expectation is almost always the same:

"If we're investing in AI, when should we expect EBITDA to improve?"

It's a fair question.

But I think it's the wrong place to start.

The biggest mistake I see companies make is expecting AI to show up in EBITDA first.

In reality, EBITDA is usually the last place you'll see the impact.

Here's why.

AI doesn't improve financial performance overnight. It changes how people work. Better ways of working improve operations. Better operations improve business performance. And eventually, those improvements appear in your financial statements.

EBITDA is a lagging indicator.

Before it moves, something else has to move first.

As a CFO, I evaluate AI investments in three stages.

Stage 1: Productivity

This is where most AI initiatives begin.

Teams save time. Reports are generated faster. Customer inquiries are answered more quickly. Marketing produces more content. Finance automates manual tasks.

These are real improvements.

But productivity alone doesn't create shareholder value.

If your team saves five hours a week and simply fills those hours with more of the same work, your EBITDA won't change.

Time saved is not the same as value created.

Stage 2: Operational Leverage

This is where AI starts changing the economics of the business.

Instead of simply working faster, your organization begins operating differently.

Sales teams spend more time selling and less time on administrative work.

Finance provides insights quickly enough to influence business decisions instead of reporting history.

Customer support resolves issues faster, improving customer retention.

Operations reduce waste, improve forecasting, and optimize inventory.

These improvements increase the capacity of the business without requiring the same level of additional resources.

That's operational leverage.

Stage 3: Financial Performance

Only after the first two stages do I expect to see AI reflected in EBITDA.

Revenue grows because sales productivity improves.

Gross margins expand because operational inefficiencies decline.

Overhead grows more slowly because repetitive work has been eliminated rather than simply accelerated.

Management allocates capital more effectively because financial information arrives faster and with greater accuracy.

Those are the drivers that improve EBITDA.

Notice what happened.

AI didn't improve EBITDA directly.

It improved the business, and the business improved EBITDA.

That's an important distinction.

Another mistake I see is expecting every AI investment to generate immediate financial returns.

Not every initiative is designed to pay back in the same reporting period.

A customer support chatbot may reduce costs within a quarter.

An AI-powered forecasting model may take much longer to influence EBITDA because its value comes from helping leadership make better strategic decisions over several years.

Both investments may be successful.

They simply create value on different timelines.

That's why I encourage CEOs to stop asking whether AI is delivering ROI.

Instead, ask whether it's improving the financial drivers that eventually determine ROI.

When I review AI investments, I ask five questions:

• Which business metric should improve?
• How will we measure it?
• When should we expect to see results?
• Which financial KPI will eventually reflect that improvement?
• When should we reasonably expect it to influence EBITDA?

Those questions create accountability.

More importantly, they separate meaningful investments from expensive experiments.

Ultimately, AI is not the investment.

Better business performance is.

Buying another AI platform doesn't create enterprise value.

Improving revenue productivity does.

Expanding gross margins does.

Increasing operating leverage does.

Making smarter capital allocation decisions does.

AI is simply one of the tools that helps make those outcomes possible.

The companies that generate the highest returns from AI won't necessarily be the ones spending the most.

They'll be the ones that understand the connection between technology, operations, and financial performance.

Because in the end, AI doesn't improve EBITDA.

Better management does.

AI simply gives great management a much more powerful advantage.

Keep growing under pressure.

-Eric

‍