Your Customers Are Very Sustainable. Until They Reach the Checkout.

Ask people whether they care about the environment and you'll get a reassuring answer.

Of course they do.

Ask whether they'd consider buying a more sustainable product.

Absolutely.

Then put two products on a shelf.

One costs CHF 12. The other costs CHF 17.

Suddenly, humanity's commitment to saving the planet becomes a little more complicated.

This isn't because people are hypocrites.

It's because people are people.

They have budgets, habits, children, deadlines and a thousand other things competing for their attention.

Now imagine asking an artificial intelligence to predict what those people will buy.

What could possibly go wrong?

The customer who never existed

In September 2026, Salesforce announced an agreement to acquire Listen Labs, an AI-powered research company.

Listen Labs combines research with customer simulations, allowing organisations to explore how people might respond to products, messages and ideas.

It's an interesting development.

Traditional market research is expensive and slow. Understanding consumers across different countries, languages and demographics takes time.

AI promises to accelerate that process dramatically.

But there's a catch.

A simulation can produce an extremely convincing explanation of consumer behaviour without proving that the behaviour will occur.

And nowhere is that distinction more interesting than in sustainability.

What tourism taught me about consumer behaviour

Earlier in my career, I worked in tourism strategy and marketing in Chile.

Tourism is a useful education in the gap between aspiration and action.

People describe extraordinary holidays they'd love to take.

They want authenticity. Adventure. Nature. Something different.

Then they confront flight prices, travel time, weather, family commitments and the practical business of actually going somewhere.

The holiday someone admires isn't necessarily the holiday they book.

The same principle applies to sustainable products.

Someone might genuinely prefer environmentally responsible packaging.

But will they pay more for it?

Will they travel to a different shop?

Will they accept an unfamiliar brand?

Will they buy it again?

Those are different questions.

Check it. 👇

We've built machines that can imitate consumer opinions. But consumer opinions were never the hardest part of understanding consumers. Behaviour was.

When the research starts agreeing with you

The danger with synthetic customers isn't necessarily that they produce ridiculous answers.

It's that they produce plausible ones.

Imagine a company developing a refillable cleaning product.

Its synthetic research panel responds positively.

Customers appreciate the environmental benefits. They understand the proposition. They express willingness to try it.

Excellent.

Then the product launches.

The refill station is inconvenient. The packaging leaks. The competitor is discounted. The customer forgets to bring the container.

None of those outcomes contradicts the customer's environmental concern.

They simply reveal that concern was never the only variable.

A synthetic customer can explain why a greener product sounds attractive. It cannot prove that someone will buy it twice.

Give synthetic research the right job

This doesn't make AI-assisted research useless.

Quite the opposite.

It can help marketers challenge assumptions, test language, explore objections and identify which hypotheses deserve investigation.

But simulated responses should not be confused with observed demand.

If the question is whether customers understand a proposition, synthetic research might help prepare the study.

If the question is whether they'll pay CHF 5 more, the answer needs stronger evidence.

Real purchasing behaviour. Pricing experiments. Returns. Repeat purchases. Customer interviews. What happens after the campaign ends.

AI can make research faster.

It cannot eliminate the difference between imagining a decision and making one.

The companies that understand that distinction will learn more quickly.

The others may discover they've developed a brilliant product for customers who don't exist.