Nobody in the automation conversation seems to be asking the most obvious question.

If AI replaces people’s jobs, how do those same people afford the products and services AI is helping businesses produce?

There’s a version of the future being sold to us where artificial intelligence handles the repetitive, the time-consuming and increasingly the skilled-but-scalable. Businesses become leaner, faster and more profitable. Work that once required entire teams can now, in some cases, be completed by a single person equipped with the right tools.

It’s a compelling vision. And in many ways, it’s already becoming reality.

But beneath the excitement surrounding AI lies a question that deserves far more attention than it’s receiving.

If large numbers of people lose income because their work is automated, who becomes the customer?

The ripple effect

Economists use the term aggregate demand to describe the total amount of spending across an economy. That spending ultimately depends on people and businesses having income to spend.

When automation displaces workers, the effects don’t stop with those individuals. The impact spreads through the economy.

Someone who loses their job spends less at restaurants. They delay replacing their car. They postpone renovating their home. They cancel subscriptions. They travel less. They spend less with local businesses.

Those businesses then earn less revenue.

In turn, they reduce marketing budgets, delay hiring, negotiate harder with suppliers and postpone investment.

The ripple doesn’t end with the person whose job disappeared. It moves through the entire economy.

History tells us that technological revolutions often create new industries and new forms of employment alongside the jobs they replace. AI may well do the same. Greater productivity, lower costs and entirely new markets could generate enormous economic value.

The uncertainty isn’t whether AI will create opportunities.

It’s whether those opportunities emerge quickly enough, and at sufficient scale, to replace the income lost during the transition.

Why this matters for South Africa

For South Africa, this question carries particular significance.

We already face one of the highest unemployment rates in the world. Millions of people are excluded from meaningful participation in the economy before widespread AI adoption has even reached full pace.

Against that backdrop, the conversation cannot simply be about efficiency.

It also has to be about participation.

Across industries, from administration and customer service to marketing, software development, accounting, consulting and creative services, businesses are exploring how AI can reduce costs and increase productivity.

That makes perfect business sense.

But if every organisation makes the same decision at roughly the same time, what happens to the broader economy that those businesses ultimately rely on?

The uncomfortable irony

The efficiency gains from AI are real.

The productivity improvements are real.

The cost savings are real.

But there’s an irony that deserves more discussion.

Businesses don’t operate in isolation.

Their customers are workers.

Their customers are other businesses.

And those businesses, in turn, rely on customers of their own.

If consumers spend less, retailers sell less.

If retailers sell less, they reduce spending elsewhere.

Marketing budgets shrink. Expansion plans are delayed. Professional services are cut. Investment slows.

The same technology that helps individual businesses become more efficient could, if adopted without sufficient economic adaptation, reduce the spending power that markets depend upon.

Whether that ultimately happens is impossible to know today.

Perhaps AI creates enough new industries, businesses and jobs to more than compensate.

Perhaps governments, educators and businesses successfully navigate the transition.

Perhaps productivity growth ultimately expands prosperity for everyone.

Or perhaps the transition proves more disruptive than many currently expect.

The point isn’t to predict the future.

It’s to ask whether we’re asking the right questions about it.

Beyond efficiency

Much of today’s conversation around AI focuses on what the technology can do.

How much faster it works.

How much cheaper it is.

How many hours it saves.

Those are important questions.

But they aren’t the only questions.

We should also be asking what kind of economy we want to build.

One where technology enhances human capability?

Or one where efficiency becomes the only measure of value?

Innovation has always changed the nature of work.

AI will almost certainly do the same.

The challenge isn’t stopping that progress.

The challenge is ensuring that technological advancement and economic participation continue to move forward together.

Because no matter how intelligent our tools become, every economy still depends on one thing.

People who can afford to participate in it.