The More Digital the Purchase, the More Strategic the Physical Store Can Become

For a long time, the evolution of digital retail seemed to follow a relatively clear direction: the consumer migrated online, e-commerce gained market share, and physical stores needed to find new reasons to remain relevant.

The answer came with omnichannel, unified commerce, ship-from-store, click & collect, and a series of initiatives created to connect what happened on the website with what happened inside the store.

Now, artificial intelligence begins to change this equation more profoundly.

We are moving towards a scenario where the consumer simply states what they need. An artificial intelligence researches, compares alternatives, considers price, availability, reputation, timing, and location, assembles a cart, and progressively may also execute the purchase.

In this model, the consumer may not need to visit the website, open the app, or even know all the retailers considered in that decision.

This changes e-commerce.

But it can also completely change the value of the physical store.

And here arises an interesting paradox: the more digital and automated the purchase becomes, the more important the physical presence of the retailer may become.

E-commerce May Be Entering an Era Without Navigation

For more than two decades, we built e-commerce around the interface.

Homepage, categories, filters, search, product page, reviews, cart, and checkout have been continuously refined to reduce the effort required to purchase.

Now, we are starting to develop technologies whose goal is precisely to eliminate part of that navigation.

Target reported that traffic generated by AI for its channels grew 2,000% in the first quarter of 2026, while AI traffic to retail websites overall advanced nearly 400%.

The company already allows consumers to discover products, build carts, and purchase through Google, Gemini, and ChatGPT, also choosing between delivery, in-store pickup, and Drive Up.

Walmart is making a similar move.

The company integrated its catalog with Gemini, allowing products to be discovered within the conversational experience itself. This integration combines purchase history, existing cart, loyalty benefits, and product availability, including hundreds of thousands of items with delivery in less than three hours and, in some cases, in as little as 30 minutes.

This may be one of the most significant turning points of this transformation.

AI is not simply creating a new storefront. It is beginning to decide which storefront makes sense.

When this happens, retail needs to stop thinking only about how to present a product to the consumer and start thinking about how to make that product eligible for an automated decision.

The New Retail Shelf is Invisible

We are already starting to see metrics for what we might call a new digital shelf: the presence of brands and products in the responses and recommendations of artificial intelligence agents.

A benchmark created by ReFiBuy in partnership with Digital Commerce 360, tracking the top 1,000 retailers, showed that 972 of the 1,000 retailers changed positions in the AI visibility ranking between the analyzed cycles.

This data is relevant because this new shelf is much more dynamic than a traditional category page.

A brand's position can change depending on the question, context, consumer intent, and the information the system can interpret.

This is where concepts like GEO, or Generative Engine Optimization, begin to emerge in an attempt to do for generative platforms what SEO did for search engines.

But perhaps the problem is even greater.

Imagine that two retailers offer exactly the same sneakers. One sells for R$ 20 less but delivers in four days. The other has the product available in a store three kilometers from the consumer and can deliver it in two hours.

Depending on that consumer's intent, proximity may be more important than price.

At this moment, something changes.

The ZIP code begins to transform into a product attribute.

And this profoundly alters the competitive logic of retail.

The Last Mile Begins to Participate in the Purchase Decision

Historically, logistics entered after the commercial decision.

First, the consumer chose the product. Then the operation needed to figure out how to deliver it.

In AI-mediated retail, these two decisions begin to converge.

Timing, availability, distance, and fulfillment capacity cease to be mere consequences of the sale and begin to influence which sale will be made.

Therefore, it may be useful to start thinking of the last mile as a kind of logistical merchandising.

Merchandising has always meant deciding which product to place in front of the consumer, in what position, and in what context.

In an AI-mediated environment, availability, proximity, and delivery capability can also help decide which product will be placed in front of that consumer.

Walmart offers a clear example.

In May 2026, the company reported that it was already offering deliveries in as little as 30 minutes in 33 U.S. markets, with over 100,000 eligible items.

The operation utilizes the proximity of stores and algorithms that consider cart size, driver availability, and distance between store and customer.

Home Depot follows a similar logic, using over 2,000 stores in the United States as fulfillment hubs to offer thousands of SKUs in as little as three hours.

These moves help show that the traditional division between store and distribution center is beginning to lose strength.

The store is not just the place where a sale happens.

It begins to participate in the digital architecture of the sale.

In Brazil, Tax Reform Makes This Equation Even More Complex

There is still a particularly important variable for Brazilian retail: tax reform.

At first glance, artificial intelligence and taxation may seem like different topics.

In practice, the two subjects begin to intersect precisely in one of the most important decisions for retail in the coming years: where to fulfill each order.

The transition to the new tax system, with the gradual replacement of current taxes by IBS and CBS, tends to increase the importance of fiscal management, the origin and destination of operations, pricing formation, and the quality of the data that supports each transaction.

For a traditional retailer, this already represents a challenge.

For an omnichannel retailer, the problem is greater.

And for an operation where an artificial intelligence can automatically choose between different inventories, stores, and logistical alternatives, the decision gains an even broader dimension.

The choice of where to fulfill an order can simultaneously involve timing, freight, availability, margin, price, and tax impact.

The inventory architecture of the future, therefore, may not be designed solely to reduce miles traveled.

It will need to optimize a much more complex economic equation.

Artificial intelligence can help find the best alternative. But this will only be possible if the retailer has sufficiently structured data to feed that decision.

It is precisely here that tax reform ceases to be an exclusively fiscal issue and begins to enter the agenda of e-commerce, supply chain, pricing, technology, and strategy.

And the location of inventory gains a new economic dimension.

The Physical Store Begins to Look Like an API

A store has an address, inventory, people, picking capacity, space, operating hours, and proximity to certain consumers.

Historically, we have viewed these characteristics as operational attributes.

In an increasingly agentic commerce, they also become digital attributes.

The store needs to continuously respond to questions such as:

Is the product available?

In what quantity?

How far from the consumer?

Can it be picked up?

Can it be separated now?

Is it possible to deliver it?

In how much time?

What is the cost of this operation?

What is the best alternative considering price, timing, and margin?

When these answers are available to the systems responsible for the purchasing journey, the store begins to function almost like a physical API of retail.

It starts to offer capacity to the digital environment.

Target is already operating on this logic at a significant scale.

In the first quarter of 2026, 79.7% of the company's merchandise sales originated from stores.

And 97.6% of sales were fulfilled by channels related to stores, including in-store purchases, shipping from stores, Order Pickup, Drive Up, and same-day delivery.

This data is important because it changes the way we view the relationship between physical and digital.

The store is not just competing with e-commerce.

It is fueling e-commerce.

Proximity Begins to Become a Product

This discussion takes on particularly interesting contours in Brazil.

We have a combination of marketplaces, Pix, WhatsApp, high smartphone penetration, distributed physical networks, and consumers accustomed to comparing price, timing, and reputation before buying.

Mercado Livre may be one of the clearest examples of how commerce, technology, and logistics can be treated as parts of a single system.

The company claims to have over 340 million users in Latin America and has positioned marketplace, logistics, technology, and financial services as integrated parts of its ecosystem for years.

In June 2026, it launched in Brazil Envios Agora, an ultra-fast delivery modality based on the seller's own inventory, with a maximum timeframe of just 25 minutes between the sale confirmation and the package dispatch.

It may seem like just a logistical improvement.

But there is a strategic change behind it.

Proximity has become a product.

And this characteristic tends to gain even more value when artificial intelligence agents begin to compare alternatives considering not only price and rating but also the possibility of receiving or picking up a product almost immediately.

The Store That Sells Less May Generate More Value

There is another important consequence of this transformation.

If a store begins to function simultaneously as a showroom, pickup point, local inventory, delivery hub, media, and source of behavioral data, evaluating its performance solely by the sales made within its four walls becomes insufficient.

Imagine two stores that generate R$ 1 million directly.

The first additionally generates another R$ 3 million in digital sales through its inventory, pickups, local deliveries, and its influence on the consumer journey.

The second does not perform any of these functions.

Financially, the two may seem similar when analyzed solely by store revenue.

Economically, they are completely different assets.

This requires a new way to measure omnichannel.

We need to assign value to proximity, availability, fulfillment capacity, experience, and the store's contribution to digital sales.

And tax reform makes this analysis even more important because the economic efficiency of this network will need to consider not only conversion and logistics but also the tax structure of each operation.

Perhaps the True Omnichannel is One Where No One Talks About Channels Anymore

For years, the goal has been to integrate store, website, app, marketplace, WhatsApp, and social media.

This made sense because the consumer clearly perceived the boundaries between these environments.

The next generation of retail may operate differently.

The consumer presents an intent.

The intelligence interprets.

The catalog responds.

The inventory informs.

The pricing competes.

The store makes available.

The logistics executes.

The payment concludes.

For the consumer, it matters little which channel performed each step.

Perhaps this is precisely the next evolution of omnichannel: to stop organizing retail around channels and start organizing it around consumer intent.

From Channel to Intent

When this happens, several traditional boundaries begin to disappear.

The physical store ceases to be the opposite of e-commerce.

The last mile ceases to be merely an operational expense.

The inventory ceases to be just a financial asset.

The catalog ceases to be just content.

Taxation ceases to be just a fiscal issue.

And artificial intelligence ceases to be just a customer service tool.

All become part of the same decision.

AI interprets. The inventory responds. The store brings closer. The logistics executes. And the economic structure defines whether that operation makes sense.

Therefore, we are facing a transformation more structural than the simple arrival of a new sales channel.

Retail is moving towards an architecture where artificial intelligences will be able to evaluate, in seconds, different combinations of product, price, inventory, location, timing, margin, taxation, and delivery.

For years, the main challenge was to bring the consumer to the product.

Now, another begins to emerge:

to ensure that the right product, in the right inventory, in the right place, with the right economic structure and the right delivery is the answer chosen by the intelligence representing that consumer.

If this scenario advances, the physical store may finally cease to be seen as a relic of traditional retail.

The more digital the purchasing decision becomes, the more relevant what is physically close to the consumer may become.

And in this new retail, the network of stores may end up revealing itself not as the past of commerce.

But as one of the most important infrastructures of its digital future.

In AI-mediated retail, proximity is no longer just logistics but becomes part of the purchasing decision.