The hardest part of AI isn’t finding a good use case. It’s deciding which opportunities are worth pursuing.

A vendor demonstrates an AI tool that can do in minutes what currently takes hours. The ROI looks compelling. Your team is enthusiastic. Someone suggests adding it to the roadmap. Why wouldn’t you?

Well, for starters, there’s the question nobody has asked yet.

Is this the best use of our resources?

Maybe the tool really is terrific. Maybe it will save hundreds of hours. But perhaps there’s another problem in the business that’s costing you millions. Or a growth opportunity that could fundamentally change your competitive position. The AI tool might be a perfectly good investment.

That doesn’t make it the right investment. And with the volume of AI ideas coming at leadership teams today, making that distinction is becoming one of the most important decisions executives face.

AI is not your business strategy

It’s a means of advancing the strategy you’ve already chosen.

That sounds obvious. But look at how AI decisions are being made in many organizations.

A vendor pitches a new capability. A competitor announces an initiative. A board member asks what the company is doing about AI.

Suddenly there’s pressure to do something. And the conversation shifts from what the business needs to accomplish to what the technology can do.

Consider two companies evaluating the same AI opportunity. One is focused on improving operating margins. The other is trying to outpace competitors through growth.

The first might use AI to reduce costs or avoid additional hiring. The second might use that same technology to increase capacity, serve more customers, or enter new markets.

Same technology. Different strategic priorities. Different measures of success. Neither approach is inherently better. The right investment depends on what the business is trying to achieve.

Before you automate something, ask whether it should exist

Peter Drucker made an observation decades ago that seems particularly relevant to AI: “There’s little value in making work more efficient if the work shouldn’t be done in the first place.”

Think about some of the things we’re rushing to automate.

Reports nobody reads. Meetings that don’t lead to decisions. Approvals that add little value. Elaborate presentations that consume hours and change nothing. AI can make all of them faster.

But why are we doing them at all? A company that automates a broken process may simply end up with a faster broken process.

Before investing in AI, leadership should understand the work itself. What purpose does it serve? What needs to change? Could the work be eliminated, simplified, or redesigned instead?

Sometimes the smartest AI decision is not to use AI.

The $1.5 million productivity gain that may not be worth $1.5 million

Even when AI improves work that genuinely matters, there’s another question worth asking: what is that improvement worth?

Consider a hypothetical company with 500 employees. An AI tool is expected to save each employee 40 hours annually.  That’s 20,000 hours. At a fully loaded labor cost of $75 per hour, the theoretical productivity value is $1.5 million. An attractive business case, right?

Maybe. But what happens to those 20,000 hours?

If employees complete their work faster, but the company doesn’t change staffing, increase output, reduce spending, or improve another measurable business outcome, it hasn’t demonstrated $1.5 million in financial value. It has released capacity.

And capacity released is not value realized.

Now suppose the company’s strategy is growth. Rather than eliminating positions, leadership uses that capacity to serve more customers without adding staff.

Or perhaps the objective is improving customer relationships, and employees spend less time on administrative work and more time with the people they serve.

Those outcomes could be far more valuable than the original labor-savings estimate. But they require deliberate choices about how the work will change. The value isn’t determined solely by what the technology saves. It’s determined by what the business does differently as a result.

Sometimes the biggest value isn’t a cost reduction

Consider what’s happening with ambient AI in healthcare.

Traditionally, physicians spend a significant portion of patient visits entering information into electronic medical records. Patients often find themselves talking to someone whose attention is divided between the conversation and a computer screen.

Ambient AI can capture and help document the conversation in the background. There’s an obvious productivity benefit. Physicians may spend less time on documentation.

But consider what else changes.

The physician can look at the patient. Watch facial expressions. Listen without constantly turning back to a screen.

The patient gets a different experience.

That may improve communication, trust, and quality of care. Depending on how the organization uses the technology, it may also create opportunities to improve physician capacity or reduce administrative burden.

The point isn’t that those outcomes are guaranteed. They need to be evaluated and measured. The point is that a business case built entirely around minutes saved could miss much of the strategic value

The most important benefit of an AI investment may have very little to do with the task being automated. A good use case can still be the wrong priority

Imagine a leadership team considering two AI opportunities.

The first automates contract review. The technology is proven, implementation looks straightforward, and the estimated return is attractive.

 The second addresses a recurring operational bottleneck that’s constraining growth. It’s more complicated. Several departments will need to change how they work. The economic potential is much larger, but so is the uncertainty.

Which one should the company pursue?

 It’s tempting to choose contract review. The problem is easy to understand, the solution is readily available, and the benefits are relatively simple to calculate.

But suppose the company’s primary objective is to accelerate growth, and the operational bottleneck is preventing it from serving additional customers.

Now the second opportunity deserves serious attention, even if it requires more work to validate.

That doesn’t automatically make it the winner. The cost, timing, organizational readiness, and confidence in the expected benefits still matter.

But those are the comparisons leadership needs to make.

Otherwise, the easiest opportunity to imagine can become the easiest one to approve.

Meanwhile, the bigger business problem remains unsolved.

What should executives know before approving an AI investment?

I’d start with five questions:

  1. Which business priority does this advance?

Not which department wants it, or which vendor is promoting it. What is the company trying to accomplish, and how will this investment help?

  1. Are we improving work that actually needs to be done?

Before automating a process, challenge the process itself. Eliminate unnecessary work. Simplify what remains. Then consider where AI can make a difference.

  1. What will change, and how will we capture the value?

 If AI releases capacity, what will happen to it? If it improves service, what business outcome should follow? If it reduces errors, what costs or risks will be reduced? A productivity estimate is not a value realization plan.

  1. How much do we trust the numbers?

 Which assumptions come from actual operating data? Which are benchmarks? Which are educated guesses? A promising opportunity with uncertain economics may deserve further validation rather than immediate approval.

  1. Is the organization prepared to deliver the result?

Do we have the right data, processes, leadership commitment, and capacity for change? A financially attractive opportunity can still be a poor near-term investment if the organization isn’t ready to execute.

Taken together, these questions help leadership distinguish between what’s possible, what’s valuable, and what’s achievable.

For private equity, the clock matters too

Private equity firms face the same choices, with an added constraint.

Time.

An AI opportunity may promise substantial productivity improvements, but those improvements don’t automatically increase EBITDA or strengthen a portfolio company’s value creation plan.

The business has to capture the benefit. And it has to do so within a timeframe that matters.

That could mean reducing external spending, avoiding planned hires, increasing profitable output, improving customer retention, or preventing margin erosion.

The right choice depends on the portfolio company’s strategy and where it sits in the investment cycle. A company preparing for an exit may favor opportunities with shorter implementation periods and more predictable economic outcomes while a company early in its holding period may have more room to pursue AI-enabled growth or larger operational changes.

Either way, a list of attractive AI use cases isn’t a value creation strategy.

The investment decisions need to reflect the business objectives, the economics, the organization’s readiness, and the time available to realize the benefit.

The real challenge is deciding what deserves your attention

Most leadership teams don’t need another list of things AI could do. They already have more ideas than they can reasonably pursue. What they need is a disciplined way to sort through those ideas, connect them to the business strategy, understand what they’re worth, and decide where to focus.

Making good AI investment decisions requires strategic alignment, economic discipline, an understanding of organizational readiness, and sometimes the conviction to say no.

AI is not a strategy. It’s a means of advancing your strategy.

And competitive advantage comes from making better business decisions, not simply adopting better technology.


This philosophy is the foundation of KAI by Kissinger Group, a 45-day executive sprint designed to help leadership teams make strategically grounded, economically sound AI investment decisions.

The financial example in this article is hypothetical and illustrates the distinction between productivity estimates and realized business value.

Meet The Author

Sondra Kiss is the co-founder and managing director of  Kissinger Group

Let’s tackle your toughest transformation challenges together! ​
Call us at (619) 743-7177 or connect with us below.
contactpage2-2.png
Follow us on
© 2024 Kissinger Group, Inc.