Shopify CEO's latest judgment: AI is advancing the shopping decision-making process, and the traditional advertising model may no longer work.

Recently, a rather interesting event has taken place in the global e-commerce industry.

Faced with Muse, the AI Agent product launched by Meta, Amazon chose to restrict its access, while Shopify took a completely opposite attitude and actively embraced it.

Both are e-commerce giants, so why do the two companies hold drastically different attitudes towards AI Agent?

Tobi Lütke, Co-founder and CEO of Shopify, believes that there are fundamental differences between the two business models behind this.

Amazon is more like a digital department store, building its competitive advantages relying on massive product inventory, traffic and transaction systems. In contrast, Shopify serves millions of independent merchants, enabling them to sell products across any channel.

For the former, AI Agent may mean losing control over the shopping entry; but for the latter, Agents may instead bring new transaction opportunities.

Recently, Tobi Lütke participated in the Sources podcast hosted by Alex Heath, and had an 84-minute interview focusing on AI-powered shopping, the business model of Agents, and the future of the entire e-commerce industry.

I think several viewpoints in this interview are well worth our attention.

First, the node at which e-commerce intervenes in users' shopping decisions has been moved forward. AI is transforming shopping from commodity consumption to intent-based consumption.

Take a home gym as an example. In the past, users might first break down the specific products they need, then search for specific items such as treadmills and purchase them one by one. Now, they will directly ask AI to buy all the things needed to build a home gym for them.

Second, Tobi also doubts the currently widely recognized business model where personal Agents draw commissions. Because the existing e-commerce ecosystem already has a complete and mature benefit distribution system, the profit margin that can be freely allocated may not be as large as people think.

Third, in the Agent era, the traditional advertising model may face a fundamental trust issue. If an AI assistant receives money from merchants before recommending products to users, does it represent the buyer or the seller after all?

Today, Silicon-based Observer will talk about how AI is changing the underlying logic of the e-commerce industry with reference to this interview.

From commodity consumption to intent-based consumption: the starting point of shopping decisions has been moved forward

Over the past two decades, the e-commerce industry has been competing around one core thing: how to make it easier for consumers to find the products they want to buy.

From the earliest product classification and keyword search, to later personalized recommendation and live-streaming sales, the essence of all these measures is to optimize the matching efficiency between products and consumers.

However, the emergence of AI Agent may fundamentally change this logic. Tobi Lütke mentioned in the interview that Shopify has observed a very interesting phenomenon.

At present, the volume of shopping transactions completed through AI Agent is still at a relatively low base, so simply discussing the growth rate does not make much sense. What is really worth paying attention to is that consumers' purchasing behavior is changing.

A typical case is building a home gym.

In the past, if a person wanted to build a gym at home, he might first search for a treadmill, then research equipment such as dumbbells, training chairs and floor mats, and finally purchase them one by one according to his budget and available space.

Throughout the process, e-commerce platforms are mainly responsible for providing product information and completing transactions. As for what to buy, how to match the equipment and how to allocate the budget, the decisions are still made by consumers themselves.

But now, some consumers have begun to put forward completely different requirements to Agents:

Buy all the things I need to build a home gym for me.

This seems to be just a change of prompt, but the underlying business logic is completely different.

Because what Agents are facing is no longer a purchase demand for a single product, but a project that requires a complete solution. They need to understand the user's budget, available space and fitness goals, judge which equipment is worth buying, and then complete product screening, matching and order placement.

Tobi even envisions that in the future, consumers may only need to upload a photo of their room or a house layout plan, and AI can participate in the entire procurement planning process.

There are many similar scenarios. For example, you want to accompany your child to complete a 3D printing project on the weekend. In the past, parents might need to research the project plan first, figure out what electronic components, magnets and printing consumables are needed, and finally place orders at different merchants respectively.

Now, you only need to tell the Agent what project you want to complete, and it can handle the full list of materials, product procurement and delivery arrangements for you.

This is an important trend Tobi has observed: More and more complex, high-unit-price purchasing behaviors are starting to be completed through Agents.

In my opinion, the truly significant change here is that AI is moving the starting point of shopping decisions from commodity demand to life demand.

In the past, there was a huge decision-making link between intent and specific commodities, which was usually completed by consumers themselves.

Now, AI has the opportunity to turn these decision-making tasks into a kind of service that can be provided at scale.

This means that the value of Agents is not only to improve the existing shopping efficiency, but also to create consumption that did not happen in the past due to excessively high decision-making costs.

Why did Amazon block Agents while Shopify took the initiative to open the door?

When consumers begin to hand over their shopping decisions to Agents, a new problem arises: 

What should e-commerce platforms that originally controlled the consumer entry do?

The different choices made by Amazon and Shopify exactly provide two very representative cases.

In Tobi's view, this divergence is not surprising at all.

Over the past two decades, the two companies have been representing two different development paths for e-commerce.

Amazon's business model is more similar to a digital department store. Consumers enter Amazon, search for products, compare prices, check reviews, and finally complete purchases. The platform attracts consumers through massive product supply, logistics fulfillment capabilities and transaction systems, and then uses traffic to help merchants complete sales.

Therefore, where consumers search, what products they see, and what they finally choose to buy are of great commercial value to the platform.

This is especially true for the advertising business. When merchants need to purchase in-platform ads to get exposure, consumers' search and browsing behaviors become important monetization resources for the platform.

But Agents may change all of this. Suppose a consumer asks AI to buy home fitness equipment for him, AI directly completes demand analysis, product comparison and procurement decisions, and finally only needs to call the e-commerce platform to complete the transaction.

In this case, many core links of traditional e-commerce platforms may be bypassed. Consumers do not necessarily need to open the e-commerce App, browse products one by one, or make decisions according to the platform's recommended order.

For platforms that rely on traffic distribution and advertising monetization, AI Agent is not just a new shopping channel, but also a potential competitor for entry control.

Of course, Tobi did not assert that Amazon restricted Muse purely to protect its advertising revenue. He also acknowledged that aggregated platforms do have stronger incentives to maintain control, and there are also problems with the current technical implementation of Agents.

But from the perspective of business models, the potential conflict between the two sides is already quite clear.

In contrast, the situation of Shopify is completely different. Shopify does not operate a large e-commerce marketplace in the traditional sense, but provides independent merchants with site building, payment, order management and sales tools.

Consumers can place orders through merchants' independent websites, and can also complete purchases through social media, offline physical stores and other channels.

Tobi has a very simple judgment on this: The more places merchants can sell their products, the greater the value of Shopify.

Therefore, no matter who places the final order, the consumer himself or the AI Agent representing the consumer, it is acceptable for Shopify. What it really cares about is whether the transaction can be completed smoothly and whether merchants can get more orders.

This is why Shopify is building more APIs and transaction systems to make it more convenient for Agents to call product and checkout capabilities.

The difference between the two business models is thus clearly reflected.

Amazon needs to consider how to maintain the relationship between consumers and the platform, while Shopify pays more attention to how to help merchants access new consumer entries.

Of course, this does not mean that Shopify will definitely become the winner in the Agent era. But at least from the perspective of business model, the conflict of interests between Shopify and Agents may be much smaller than that of traditional traffic-based platforms.

In my opinion, this is also the key to understanding the next round of e-commerce competition.

When AI begins to control shopping decisions, the competitive focus of e-commerce platforms may shift from how to keep users on their own platforms to how to make their own products and services prioritized by Agents.

Agents want to draw 2% commission from each transaction, but who will pay for it?

If it is already clearly visible that AI is changing the way people shop, then another more realistic question arises: 

How on earth should Agents make profits?

In this interview, host Alex Heath mentioned a common idea shared by Silicon Valley Agent startups at present.

Many entrepreneurs believe that in the future, personal Agents can build their business models through transaction commission. For example, if AI helps a consumer complete an order of 100 US dollars and draws a 2% commission, it can get 2 US dollars in revenue.

At first glance, this model seems quite reasonable.

After all, Agents help users complete product research, comparison and procurement, creating actual value, so it makes sense for them to get a certain percentage of transaction revenue.

But Tobi directly raised a question: Who on earth should pay this 2% commission?

There is a reality that is easy to ignore here. E-commerce transactions already have a complete and mature benefit distribution system: participants such as payment institutions, credit card networks and e-commerce service providers all obtain corresponding revenue from transactions.

According to Tobi's disclosure, a large part of Shopify's approximately 3% transaction fees actually goes to credit card companies. That is to say, although it seems that there are several percentage points of transaction revenue, the profit margin that can be freely allocated may not be that large.

Now that Agents suddenly join the transaction chain and want to take 2% of the share, the problem becomes complicated.

If merchants are required to bear the cost, it will further increase their customer acquisition cost. If the platform is required to bear it, it means that existing participants need to give up part of their own profits. If consumers are required to bear it, Agents must prove that the value they provide is sufficient to justify users paying extra fees.

Therefore, Tobi is in no hurry to judge that commission drawing will definitely become the final business model of Agents.

In his opinion, the more important question is how much value Agents can create for users.

For example, an Agent can help consumers compare the fee plans of different banks, find unreasonable long-term payment charges, and even automatically complete service switching. If it can help users save hundreds of dollars every year, consumers have every reason to pay a certain fee for it.

Under this model, Agents do not necessarily need to participate in commission drawing from commodity transactions, but can charge users directly. In the future, pay-as-you-go, subscription fees, or transaction-related hybrid models may also emerge.

But what is more worth discussing than the business model is another question raised by Tobi: When Agents truly make decisions on behalf of consumers, the traditional advertising model may encounter a trust crisis.

Suppose you have a personal shopping assistant that knows you very well, knows your budget, living habits and purchasing preferences, can help you compare different brands and choose the most cost-effective products.

But one day, you find that it recommends a certain product not because it is the most suitable for you, but because the merchant paid it advertising fees.

Then, will you still trust it?

In Tobi's view, this constitutes a breach of user trust.

The traditional advertising model is mainly on the seller's side. Merchants pay for exposure, hoping to influence consumers' purchasing decisions. But the value of personal Agents is precisely built on the premise that they represent the interests of the buyer.

If an Agent needs to represent consumers to choose the most suitable products, and also determines the recommendation order according to the fees paid by merchants, a conflict of interest may arise between the two.

This is also why the advertising business model in the Agent era may need to be redesigned.

However, Tobi does not believe that advertising will disappear as a result.

No matter how powerful AI is, new brands still need to reach consumers, and new products still need to gain exposure.

In the past, Facebook and Google's advertising systems helped a large number of small merchants find their first batch of users. In the future, the Agent world will also need a certain mechanism for product discovery and supply-demand matching.

However, this mechanism must answer a new question: How to help merchants obtain traffic without damaging consumers' trust in AI assistants?

This may be one of the most noteworthy directions for Agent commercialization. Whoever can strike a balance between sellers' customer acquisition needs and buyers' decision-making trust will likely be able to build a new generation of commercial distribution platforms.

This article is from the WeChat official account "Silicon-based Observation Pro", author: Silicon-based Observer, published with authorization from 36Kr.