Imagine starting a new online perfumery today. You have a good network within the fragrance industry, access to interesting brands and a solid understanding of the market. But instead of developing a webshop, you decide to organise the business differently.
You build a commercial infrastructure that allows AI assistants to communicate directly with your systems. Your assortment, product information, prices and availability are accessible through technical connections. A consumer looking for a fragrance can discover and compare your products within their own AI environment and complete the purchase there as well. The order is then processed directly by your system and fulfilled by an external logistics partner. The consumer has never visited your website, simply because there isn't one.
Your organisation consists of a small team responsible for brand selection, supplier relationships, commercial agreements and overseeing the quality of service. AI supports operational processes, from stock planning and order processing to customer service. Consumers mainly interact with your business through external AI platforms, while your own organisation handles the commercial operations.
Technology is moving increasingly in this direction. AI platforms can already process structured product information, and systems are emerging that allow consumers to select and order products without visiting a traditional webshop. The technical foundations for such a model are taking shape, although its commercial viability at scale has yet to be demonstrated. It does, however, offer a different way of thinking about online retail.
A Different Way to Build a Retail Business
For decades, online retail has been organised around attracting visitors to a retailer's own website. Retailers invest in search engine optimisation, advertising, product presentation, usability and conversion. The website serves both as the shop and as the environment in which the retailer builds its relationship with consumers. The economics follow the same model: visitors are brought into the retailer's own sales environment, where the purchase is expected to take place.
When consumers use AI assistants to discover, compare and order products, that structure changes. Someone might ask for an interesting woody fragrance within a certain budget. The assistant compares available products, provides information about the fragrances and checks where they can be purchased. The consumer makes a choice and asks the assistant to place the order, without visiting different webshops.
For existing retailers, this means adapting their current infrastructure to a different way of discovering and purchasing products. A new retailer can build its business around this development from the beginning. There is no existing digital storefront to maintain and no previous investment in one to recover. The interesting part is not only how technology can make retailers more efficient, but also the kind of business it allows them to create. A business that is no longer organised around attracting visitors, but around making products available and fulfilling orders.
A Different Allocation of Investment
A traditional online perfumery invests in its webshop, marketing, systems, staff and logistics. Even when certain activities are outsourced, a significant part of the commercial organisation remains focused on attracting and serving visitors within its own environment.
An AI-native retailer can allocate its resources differently. Product information and commercial connections with AI platforms form the basis of sales. AI supports internal processes, while a third-party logistics provider handles storage and delivery. Investments that would normally go into webshop development, conversion optimisation or certain marketing activities can be directed towards other parts of the business, such as the assortment, stock financing, supplier relationships and customer service.
This does not mean that costs disappear. Good product information requires maintenance, technical connections need to work reliably, and access to consumers through AI platforms may come at a commercial cost of its own. But the allocation of costs and capital changes. And that can allow a smaller business to compete in a different way.
Established retailers retain important advantages. Purchasing power, brand recognition, a large customer base and established commercial relationships do not disappear because consumers start using AI. But investments in a retailer's own digital storefront provide less differentiation when that environment is no longer an essential part of the purchase. For a new retailer, the calculation can therefore be different. Not because scale loses its value, but because part of the infrastructure previously needed to sell at scale now sits outside the business itself.
Who Decides Where the Consumer Buys?
When an AI assistant recommends products, the platform influences which retailers become visible. The selection of products and sellers takes place without consumers visiting different stores themselves. This creates a new commercial layer between retailer and consumer. That dependency is not entirely new. Google, marketplaces and social platforms have long influenced how consumers discover retailers. But an AI assistant goes further by not only presenting sellers, but also guiding the comparison and purchase within the same environment.
This does not automatically weaken the relationship between consumer and retailer. Imagine someone buying a fragrance through AI from a specialist perfumery and being pleased with the delivery and service. Next time, they simply ask their assistant to order from the same perfumery again. There is no need to visit its website. If the assistant has access to the purchase history and respects the consumer's preference, a positive experience can lead directly to another order.
There is also a vulnerability here. The consumer may remain loyal to a retailer, while the information needed to turn that preference into another purchase is largely held by the AI platform. Consumer preference and control over the customer relationship do not necessarily sit with the same party. For retailers, this creates a different form of dependency, but also another way to serve returning customers. The question is how much influence they retain over that relationship when direct interaction takes place through an external assistant.
The Other Side of Selective Distribution
For niche and luxury fragrance, another issue arises. A retailer can organise its entire commercial infrastructure around AI, but that does not automatically give it access to the brands it wants to sell. Within selective distribution, fragrance brands decide which retailers they work with and what requirements they place on the selling environment. Product presentation, the appearance of the store and the quality of advice can all be part of those conditions. A retailer without its own webshop does not necessarily fit into this established model. After all, its products are presented within an AI environment over which the retailer itself has limited control.
But selective distribution works both ways. Brands select their retailers, while retailers decide which brands to carry.
The market now has a large and growing number of fragrance brands. Many are looking for access to consumers and competing for attention within a limited number of retail channels. This gives an AI-native retailer an interesting choice. It can try to meet the existing distribution requirements of brands that insist on a traditional selling environment, or it can work with brands that are open to a different way of presenting and selling their products. That does not have to mean choosing between strong brands and inferior alternatives. There are high-quality independent brands that are still relatively unknown. For these brands, access to new forms of product discovery can be attractive, particularly when consumers are not searching exclusively for familiar brand names.
This brings us to a distinction that could become increasingly important in the fragrance industry. A consumer looking specifically for a fragrance from a particular house will not necessarily accept an alternative. Brands with strong consumer demand of their own therefore have a different negotiating position from those that depend heavily on discovery for their growth. But when a consumer asks an AI assistant for a particular type of fragrance, with certain scent characteristics, within a budget and for a specific occasion, the situation is different. The assistant can suggest several brands that match those preferences. A brand that is not available through the relevant retail infrastructure may be absent from that particular comparison, while another brand is recommended instead.
This creates a new consideration for both brands and retailers. The brand wants to maintain control over how its products are offered. The AI-native retailer wants to develop an assortment that fits its commercial infrastructure and the consumers it can reach. Both interests are understandable, but they do not always align. For brands with sufficient consumer demand of their own, maintaining strict distribution conditions can remain a sensible choice. For smaller or less established houses, the same decision can mean missing certain opportunities for product discovery and sales. And for the retailer, the question becomes which brands it really needs to create an attractive assortment. When enough alternatives are available, it may decide not to invest in meeting the requirements of particular brands.
This is not the end of selective distribution. But it can change the relationship between brands and retailers, especially if new commercial models manage to reach a meaningful number of consumers.
What Remains of the Store?
Fragrance remains a sensory product. Scent, emotion, discovery and trust all play an important role, particularly within niche and luxury fragrance. An AI assistant can process information and recognise preferences, but it cannot let someone actually smell a fragrance. Yet discovery and purchase do not have to take place in the same environment. Consumers already discover fragrances through subscription services, carefully curated fragrance selections and physical brand boutiques, while a later purchase may take place through another channel. When AI also supports selection and ordering, that separation can develop further.
This means different parties can take on separate roles within the commercial process. One introduces consumers to new fragrances, another provides the physical brand experience, and a third handles the transaction and delivery. It also changes how the value of these different contributions is assessed. The party investing in discovery or brand experience is not necessarily the one making the final sale.
Physical perfumeries, brand boutiques and traditional webshops therefore remain relevant. Their role may change, but their contribution to market development and brand experience does not disappear. Existing retailers can also adapt their organisations and use the same technology. The new AI-native retailer mainly shows that a business no longer necessarily needs to develop its own digital storefront to play a commercial role. The value of its assortment, supplier relationships and service remains, while access to consumers and the execution of purchases can increasingly take place through other systems.
At first sight, this appears to be mainly a technological change in online retail. But when such a model becomes commercially attractive, it also affects the way brands organise their distribution. Brands can continue to decide under which conditions their products are sold. At the same time, retailers gain new possibilities to organise their businesses and assortments around brands that fit their commercial model.
The underlying question then becomes how long all brands can continue to apply the same distribution conditions as the way consumers discover and buy products changes.
You build a commercial infrastructure that allows AI assistants to communicate directly with your systems. Your assortment, product information, prices and availability are accessible through technical connections. A consumer looking for a fragrance can discover and compare your products within their own AI environment and complete the purchase there as well. The order is then processed directly by your system and fulfilled by an external logistics partner. The consumer has never visited your website, simply because there isn't one.
Your organisation consists of a small team responsible for brand selection, supplier relationships, commercial agreements and overseeing the quality of service. AI supports operational processes, from stock planning and order processing to customer service. Consumers mainly interact with your business through external AI platforms, while your own organisation handles the commercial operations.
Technology is moving increasingly in this direction. AI platforms can already process structured product information, and systems are emerging that allow consumers to select and order products without visiting a traditional webshop. The technical foundations for such a model are taking shape, although its commercial viability at scale has yet to be demonstrated. It does, however, offer a different way of thinking about online retail.
A Different Way to Build a Retail Business
For decades, online retail has been organised around attracting visitors to a retailer's own website. Retailers invest in search engine optimisation, advertising, product presentation, usability and conversion. The website serves both as the shop and as the environment in which the retailer builds its relationship with consumers. The economics follow the same model: visitors are brought into the retailer's own sales environment, where the purchase is expected to take place.
When consumers use AI assistants to discover, compare and order products, that structure changes. Someone might ask for an interesting woody fragrance within a certain budget. The assistant compares available products, provides information about the fragrances and checks where they can be purchased. The consumer makes a choice and asks the assistant to place the order, without visiting different webshops.
For existing retailers, this means adapting their current infrastructure to a different way of discovering and purchasing products. A new retailer can build its business around this development from the beginning. There is no existing digital storefront to maintain and no previous investment in one to recover. The interesting part is not only how technology can make retailers more efficient, but also the kind of business it allows them to create. A business that is no longer organised around attracting visitors, but around making products available and fulfilling orders.
A Different Allocation of Investment
A traditional online perfumery invests in its webshop, marketing, systems, staff and logistics. Even when certain activities are outsourced, a significant part of the commercial organisation remains focused on attracting and serving visitors within its own environment.
An AI-native retailer can allocate its resources differently. Product information and commercial connections with AI platforms form the basis of sales. AI supports internal processes, while a third-party logistics provider handles storage and delivery. Investments that would normally go into webshop development, conversion optimisation or certain marketing activities can be directed towards other parts of the business, such as the assortment, stock financing, supplier relationships and customer service.
This does not mean that costs disappear. Good product information requires maintenance, technical connections need to work reliably, and access to consumers through AI platforms may come at a commercial cost of its own. But the allocation of costs and capital changes. And that can allow a smaller business to compete in a different way.
Established retailers retain important advantages. Purchasing power, brand recognition, a large customer base and established commercial relationships do not disappear because consumers start using AI. But investments in a retailer's own digital storefront provide less differentiation when that environment is no longer an essential part of the purchase. For a new retailer, the calculation can therefore be different. Not because scale loses its value, but because part of the infrastructure previously needed to sell at scale now sits outside the business itself.
Who Decides Where the Consumer Buys?
When an AI assistant recommends products, the platform influences which retailers become visible. The selection of products and sellers takes place without consumers visiting different stores themselves. This creates a new commercial layer between retailer and consumer. That dependency is not entirely new. Google, marketplaces and social platforms have long influenced how consumers discover retailers. But an AI assistant goes further by not only presenting sellers, but also guiding the comparison and purchase within the same environment.
This does not automatically weaken the relationship between consumer and retailer. Imagine someone buying a fragrance through AI from a specialist perfumery and being pleased with the delivery and service. Next time, they simply ask their assistant to order from the same perfumery again. There is no need to visit its website. If the assistant has access to the purchase history and respects the consumer's preference, a positive experience can lead directly to another order.
There is also a vulnerability here. The consumer may remain loyal to a retailer, while the information needed to turn that preference into another purchase is largely held by the AI platform. Consumer preference and control over the customer relationship do not necessarily sit with the same party. For retailers, this creates a different form of dependency, but also another way to serve returning customers. The question is how much influence they retain over that relationship when direct interaction takes place through an external assistant.
The Other Side of Selective Distribution
For niche and luxury fragrance, another issue arises. A retailer can organise its entire commercial infrastructure around AI, but that does not automatically give it access to the brands it wants to sell. Within selective distribution, fragrance brands decide which retailers they work with and what requirements they place on the selling environment. Product presentation, the appearance of the store and the quality of advice can all be part of those conditions. A retailer without its own webshop does not necessarily fit into this established model. After all, its products are presented within an AI environment over which the retailer itself has limited control.
But selective distribution works both ways. Brands select their retailers, while retailers decide which brands to carry.
The market now has a large and growing number of fragrance brands. Many are looking for access to consumers and competing for attention within a limited number of retail channels. This gives an AI-native retailer an interesting choice. It can try to meet the existing distribution requirements of brands that insist on a traditional selling environment, or it can work with brands that are open to a different way of presenting and selling their products. That does not have to mean choosing between strong brands and inferior alternatives. There are high-quality independent brands that are still relatively unknown. For these brands, access to new forms of product discovery can be attractive, particularly when consumers are not searching exclusively for familiar brand names.
This brings us to a distinction that could become increasingly important in the fragrance industry. A consumer looking specifically for a fragrance from a particular house will not necessarily accept an alternative. Brands with strong consumer demand of their own therefore have a different negotiating position from those that depend heavily on discovery for their growth. But when a consumer asks an AI assistant for a particular type of fragrance, with certain scent characteristics, within a budget and for a specific occasion, the situation is different. The assistant can suggest several brands that match those preferences. A brand that is not available through the relevant retail infrastructure may be absent from that particular comparison, while another brand is recommended instead.
This creates a new consideration for both brands and retailers. The brand wants to maintain control over how its products are offered. The AI-native retailer wants to develop an assortment that fits its commercial infrastructure and the consumers it can reach. Both interests are understandable, but they do not always align. For brands with sufficient consumer demand of their own, maintaining strict distribution conditions can remain a sensible choice. For smaller or less established houses, the same decision can mean missing certain opportunities for product discovery and sales. And for the retailer, the question becomes which brands it really needs to create an attractive assortment. When enough alternatives are available, it may decide not to invest in meeting the requirements of particular brands.
This is not the end of selective distribution. But it can change the relationship between brands and retailers, especially if new commercial models manage to reach a meaningful number of consumers.
What Remains of the Store?
Fragrance remains a sensory product. Scent, emotion, discovery and trust all play an important role, particularly within niche and luxury fragrance. An AI assistant can process information and recognise preferences, but it cannot let someone actually smell a fragrance. Yet discovery and purchase do not have to take place in the same environment. Consumers already discover fragrances through subscription services, carefully curated fragrance selections and physical brand boutiques, while a later purchase may take place through another channel. When AI also supports selection and ordering, that separation can develop further.
This means different parties can take on separate roles within the commercial process. One introduces consumers to new fragrances, another provides the physical brand experience, and a third handles the transaction and delivery. It also changes how the value of these different contributions is assessed. The party investing in discovery or brand experience is not necessarily the one making the final sale.
Physical perfumeries, brand boutiques and traditional webshops therefore remain relevant. Their role may change, but their contribution to market development and brand experience does not disappear. Existing retailers can also adapt their organisations and use the same technology. The new AI-native retailer mainly shows that a business no longer necessarily needs to develop its own digital storefront to play a commercial role. The value of its assortment, supplier relationships and service remains, while access to consumers and the execution of purchases can increasingly take place through other systems.
At first sight, this appears to be mainly a technological change in online retail. But when such a model becomes commercially attractive, it also affects the way brands organise their distribution. Brands can continue to decide under which conditions their products are sold. At the same time, retailers gain new possibilities to organise their businesses and assortments around brands that fit their commercial model.
The underlying question then becomes how long all brands can continue to apply the same distribution conditions as the way consumers discover and buy products changes.