When Trust Becomes Machine-Readable, What’s the future for Black Friday & eCommerce?

Consumers Don’t Want AI To Shop For Them

At Least Not Yet.

Much of the conversation around AI commerce focuses on agents buying products autonomously.

The data suggests that consumers are comfortable using AI to compare products.

They’re comfortable using AI to check value.

They’re comfortable using AI to help determine whether a promotion is genuine or if there are anti-trust signals / further alternatives.

What they’re significantly less comfortable with is allowing AI to complete the purchase itself and that’s an important distinction, because it suggests consumers don’t currently see AI as a salesperson, they see it as an auditor, a verifier, a second opinion, a referee.

That’s an important consideration.

For as long as I can remember, marketing has largely been about persuasion.

AI introduces the possibility of immediate, independent verification.

That changes the rules.

The Real Threat Isn’t AI Buying Products.

It’s AI Judging Brands.

Historically, retailers controlled much of the information consumers used to make decisions.

Brands controlled the message.

Brands controlled the narrative.

Brands controlled how value was presented.

Increasingly, consumers are asking AI the questions they previously asked retailers directly.

Is this actually a good deal?

Is this product worth the money?

Can I trust these reviews?

Is this retailer reliable?

The interesting thing is that AI evaluates those questions differently from humans. Humans can be persuaded but AI only really cares about what it sees as signals or evidence.

• Structured product data.
• Review quality.
• Price history.
• Delivery performance.
• Warranty information.
• Customer outcomes.

In other words, proof.

This is why I think the biggest shift coming isn’t AI-generated shopping recommendations. It’s the possibility that trust itself becomes machine-readable.

Because when that happens, brands stop competing solely for attention.

They start competing for validation, verification and consistent trust.

Musing about the future.

This is where Meta’s Muse becomes genuinely interesting.

Not because it changes ecommerce tomorrow, not because it suddenly enables autonomous shopping and certainly not because it is reportedly outsourcing some of its automated calling features to human agents.

But because of what it tells us about where AI is heading. The key here is not where it is today within it’s development cycle, it’s the intent beyond the learning phase.

Most of today’s AI systems are answer engines. You ask, it tells and you add nuance from there. ChatGPT.

Muse represents a fundamentally different line of thinking.

Meta describes Muse as a “world model” capable of understanding and predicting how environments evolve over time.

Today, much of its application sits inside simulations, gaming and virtual environments, but that’s almost beside the point. It’s still a product in development and should not be judged without that caveat.

The important question isn’t what Muse does today. It’s what the development of world models tells us about where AI research is investing its energy.

Because prediction has always been the holy grail of commerce.

We want to know what customers will do tomorrow, but we are often tied to insights based on historic data.

That’s why entire industries have been built around forecasting, propensity modelling, customer segmentation, recommendation engines and predictive analytics.

In other words, commerce has spent decades trying to infer intent from historical signals.

World models hint at a future where AI becomes better at understanding context, anticipating behaviour and recognising emerging intent.

I want to be clear that we’re not yet near a world where AI understands human motivation better than humans themselves. But if these world models continue to improve, the question eventually shifts from:

“What is this customer searching for?”

to:

“What is this customer likely to need, AND need next?”

That’s a much more consequential question.

If AI increasingly mediates discovery, marketers face an entirely new challenge:

How do you become the recommendation before the customer starts searching?

The Real question For Retailers

The biggest misconception in ecommerce today is that AI is simply a new channel.

It’s a complete shift.

The YouGov data hints at it. Meta’s Muse hints at it.

In a future where AI helps customers determine what is true, the brands that win will be the ones whose value is easiest to understand, validate and trust, by both humans and machines.

So, What Happens Next?

Will AI change Black Friday forever? Probably.

The more interesting question is whether Black Friday itself is a symptom of an inefficient market that is already in the process of change.

After all, Black Friday exists because consumers don’t have perfect information and it acts as a concentrated solution that purports to solve that problem, when it exacerbates it:

• They don’t know if they’re getting the best price – but are made to decide quickly
• They don’t know if a discount is genuine – but are made to decide quickly
• They don’t know which retailer is most trustworthy – but are made to decide quickly

AI is starting to address all three. If AI becomes increasingly capable of validating value, understanding intent and recommending action, then the balance of power shifts from persuasion to easily accessible, trustable proof.

This is why we think the most important question raised by this year’s research isn’t how consumers will use AI. It’s how retailers remain relevant when customers increasingly trust the intelligence sitting beside them, more than the marketing sitting in front of them.

If trust becomes machine-readable, the future of ecommerce won’t be won by the brands that shout the loudest.

It will be won by the brands that can prove they’re worth listening to.