Commerce has spent the last twenty years optimizing for discovery.

Search became infrastructure.

Reviews became currency.

Marketplaces became trust layers.

Brands invested heavily to be found, understood, and chosen.

That model assumed one thing.

The customer remained in direct control of the process.

They searched, compared, decided.

What stayed with me from our conversation with Berengere Chaintreau-Fuchs was the possibility that this sequence is beginning to change.

Not because consumers are buying through AI.

That remains limited.

But AI is increasingly shaping what consumers see before they ever arrive at a product, retailer, or marketplace.

The point matters.

Discovery has always been one of the most expensive parts of commerce.

If the layer controlling discovery changes, the economics underneath it may change as well.

The question is how much trust consumers are willing to delegate to that new layer.

One reason the answer remains unclear is that trust in commerce has never been universal.

It has always been contextual.

Consumers trust Amazon differently than they trust an independent retailer.

They trust buying batteries differently than booking a holiday.

They trust routine purchases differently than emotional ones.

Berengere made this point clearly.

That matters because much of the current conversation around agentic commerce treats shopping as if it were a friction problem waiting to be solved.

Find, compare, buy faster.

The assumption underneath it is that convenience naturally compounds into adoption.

History offers reasons to be careful with that assumption.

E-commerce itself took years to build trust around payments.

Marketplaces took years to build trust between strangers.

Digital wallets took years to become routine.

The progression was gradual because trust compounds more slowly than technology.

AI may now be entering the same cycle.

Consumers appear increasingly willing to use it for information.

Delegating decisions is a different threshold.

Berengere used an expression that stayed with me:

Areas where the amount of available information has already exceeded the consumer’s ability to process it properly.

That observation says something important about where AI may find its first real footing in commerce.

The assumption in many current discussions is that AI will transform shopping by replacing existing behavior. The evidence so far suggests something narrower. It is finding its place where decision fatigue already exists.

Insurance is one example.

Financial products are another.

Even something as mundane as finding the right battery, understanding compatibility, or comparing hundreds of nearly identical options creates friction that most consumers would happily reduce.

These are not high-emotion purchases. They are high-friction purchases.

That difference shapes the opportunity.

In categories where the act of discovery has little intrinsic value, reducing cognitive effort carries obvious appeal. The role of AI becomes less about persuasion and more about compression. Less time spent filtering. Fewer wrong turns. Faster clarity.

Berengere made a related point during the discussion. Online commerce has spent years solving transactions. It has spent far less time solving decision overload.

That may explain where adoption moves first.

Not where shopping is enjoyable.

Where shopping is exhausting.

Reducing decision fatigue requires a source of confidence.

That is where the economics become more interesting.

For most of the internet era, brands built trust inside their own environments. A website, a product page, paid acquisition, conversion optimization. The transaction happened there, and much of the persuasion happened there as well.

AI shifts part of that process outward.

Before a consumer reaches a product page, the system may already have interpreted the product, the brand, and the surrounding market context.

That interpretation is not built from one source.

Berengere pointed to something important here. Reviews, Reddit discussions, external mentions, Trustpilot scores, and broader social proof are increasingly part of how AI forms an understanding of a company or product.

This changes how trust compounds.

A strong website still matters.

Product quality still matters.

But external credibility begins to carry more weight because it becomes part of the discovery layer itself.

That creates a structural shift.

For years, smaller brands could compete by being more efficient at acquisition. Better ads. Sharper positioning. Better funnels.

If AI increasingly narrows the field before discovery even begins, reputation starts to influence visibility much earlier in the process.

That does not eliminate smaller brands.

It changes what they must build first.

The question becomes less about traffic.

And more about whether enough trusted signals exist for AI to surface them at all.

If AI begins to influence discovery before a consumer reaches a retailer, the obvious response would be for retailers to strengthen the environments they already control.

That was one of the strongest threads in Berengere’s thinking.

Much of the discussion around agentic commerce assumes the transaction will eventually move outside the retailer’s ecosystem. Through ChatGPT, voice agents, or external interfaces. The near-term evidence points elsewhere.

Amazon’s own data around Rufus is one example. Customers who interact with Rufus show materially higher purchase intent, but the interaction still happens inside Amazon’s environment. The trust layer, payment layer, delivery layer, and return logic remain intact.

That changes the strategic implication.

Retailers may not need to surrender commerce to external agents. They may instead need to rebuild discovery inside their own ecosystems.

Search becomes conversational.

Personalization becomes contextual.

Navigation becomes adaptive.

The objective stays the same.

Reduce the effort required to reach a confident decision.

That is where trust compounds fastest.

Not when a customer is pushed into a new environment.

When a familiar environment becomes more intelligent.

Much of the public conversation around agentic commerce still centers on autonomy.

AI buying on behalf of consumers.

AI managing recurring purchases.

AI making decisions without supervision.

The infrastructure may arrive before the behavior does.

That possibility came up repeatedly during the discussion.

Consumers are already using AI at scale, often without thinking about it. Google summaries are one example. Product recommendations are another. Search itself is becoming increasingly mediated.

The behavioral shift is already underway.

The difference is that most of it still stops short of delegation.

Consumers appear comfortable with AI reducing complexity.

They remain more cautious when AI starts carrying responsibility.

That gap may define the next phase of commerce.

The companies building for full autonomy may be directionally right.

The sequence may simply be different.

First AI becomes infrastructure.

Then it becomes habit.

Only later might it become trusted enough to act alone.

By then, the companies that understood where trust begins may be better positioned than those who only focused on what the technology could do.

EVOLVE is a private community of senior commerce leaders, founders, operators, investors, advisors, and board members across nearly 50 markets.

We bring together VC’s, PE’s, Family offices, board-level leaders, ambitious founders, and senior operators across commerce markets through ongoing structured conversations.

The community exists to accelerate access to people, perspectives, and opportunities that would be difficult to create alone.

Every week, members engage with experienced practitioners building, investing in, and scaling companies across commerce, technology, retail, marketplaces, payments, and adjacent industries.

If you believe meaningful opportunities are often created through trusted relationships, shared experience, and long-term thinking, we'd be glad to hear from you.

The room stays small by design.

The conversations stay practical for the same reason.

That is EVOLVE

If you’d like to learn more how to partner with, please email us at
[email protected] or [email protected]