In the past two years, there has been a lot of talk about AI. Many companies, when they think of AI, immediately think of writing copy, making posters, generating short video scripts, or helping customer service reply to a few messages. This direction is certainly useful, at least it can save some labor and produce content faster. But if you only look at this, it's easy to underestimate AI's impact on the consumer industry. What I'm more concerned about is another thing: AI has begun to enter the steps before consumers place an order. In the past, consumers had to search, browse, compare, judge, and finally place an order themselves. Now, some of these steps are starting to be handed over to AI. It can understand a person's needs, help filter products, help combine orders, and even continue to handle payment and fulfillment. This will have a profound impact on consumer goods, retail, platforms, and channels. One sentence to order coffee, behind it is not just a few fewer taps on the phone Let's look at a small scenario first. In the past, if a person wanted to order a cup of coffee, they would go through several steps: open the food delivery app, search for coffee, click into the store, look at the price, delivery fee, and discounts, then choose the size, temperature, sugar level, and finally place the order. Everyone is familiar with this set of actions. So familiar that we don't even think it's troublesome. But with AI, the path changes. The user might just say: "Order me an iced Americano that can be delivered nearby as soon as possible." Then, AI understands the need, judges nearby merchants, compares prices, delivery times, and user preferences, generates an order, and waits for user confirmation. On January 15 this year, the Qianwen App announced integration with Alibaba businesses such as Taobao, Alipay, Taobao Flash Purchase, Fliggy, and Amap, enabling functions like ordering food, buying things, and booking flights, and opened testing to users. At the press conference, the demonstration was "Order 40 cups of Chagee's Boya Juexian," where Qianwen called Taobao Flash Purchase to complete the order and paid via Alipay's AI Pay. The point here is not that food delivery has a new entrance. What's really worth noting is that AI has begun to enter real transactions. Users are no longer just asking "What's good to eat nearby?" and then searching in the app themselves. It can continue further, directly entering ordering, generating orders, and making payments. Once this step is proven, the entrance to the consumer industry will change. In the past, people searched for products; now, needs come first In the past, the consumer goods industry had a clear chain. Brands advertise, platforms provide traffic, retailers stock shelves, and consumers see the products, then search, compare, and order. So in the past, what many companies competed for was also clear: whoever can be seen, remembered, ranked high in search, and occupy shelf space is more likely to be bought. After e-commerce emerged, shelves moved from offline to online. After content platforms rose, shelves moved from search pages to feeds. But the underlying action hasn't changed much. Consumers still have to find products themselves. With AI, consumers may not necessarily mention the brand first, nor the category. They might directly express a life need. For example, they don't say "I want to buy C&S tissue," but rather: "There are elderly and children at home. The tissue should be soft, not shed lint, and not too expensive. Help me buy a case." Is there a brand in this sentence? No. Is there a category? Yes, but not complete. What's truly valuable are the conditions behind: elderly, children, soft, no lint, not too expensive, a case. What AI needs to do is translate these words into product choices. At this point, the problem brands face changes. In the past, you had to make consumers remember you; now, you also have to make AI know: in which scenarios should you be included in the candidate list. This change is subtle but crucial. Shelves will change from fixed positions to temporary generation In the past, when we talked about shelves, they were basically fixed.

Offline stores have end caps, stack displays, and positions near the checkout counter;

E-commerce has search result pages, homepage recommendations, and event venues;

Content platforms have feeds, influencer seeding, and live streams. These are all visible positions. The shelves brought by AI are different. They are not necessarily a fixed page, nor do they look the same for everyone. For the same tissue, if one user says "There's a baby at home," another says "Office procurement," and a third says "For the elderly, not too hard," the choices AI generates may differ. This means that in the future, a brand will not just compete for a single shelf position. It needs to enter many specific scenarios: Family packs, baby use, elderly use, office procurement, high cost-performance, gifting, stockpiling, and instant replenishment. In the past, a brand saying "Family first choice" sounded fine. But with AI, this is not enough. AI needs clearer information: how many plies, how many sheets, whether it's scented, who it's suitable for, the cost per sheet, whether reviews mention lint, whether there's nearby inventory, and how soon it can be delivered. In the past, many brand expressions were for people to see. In the future, some information also needs to be machine-readable. This is not something that can be solved by writing a few nice introductions. Product information, user reviews, price explanations, scenario tags, and fulfillment data all need to be clearer. Instant retail will be rewritten first Why will this change happen first in food delivery, coffee, milk tea, and instant retail? The reason is simple: these needs are frequent enough, and the decision cost is low. Hungry, order food;

Thirsty, order milk tea;

Out of tissue at home, buy a case;

Child has a spring outing tomorrow, buy some snacks on the spot;

Friends come over, replenish a few drinks. These needs are naturally suitable for expression in one sentence. If it's buying a house, a car, or insurance, users are unlikely to fully delegate the decision to AI. But for a cup of coffee, a meal delivery, a case of tissue, or a bag of rice, users are willing to try. Instant retail also has a characteristic: it has ready-made fulfillment capabilities. Whether there's stock nearby, how soon it can be delivered, and the delivery fee are data the platform already has. AI just connects "user expressing needs" and "platform organizing orders." Meituan is also making similar attempts. In September 2025, Meituan's first AI Agent product "Xiaomei" launched its public beta, positioned as a life assistant. Public reports mentioned it can complete food delivery orders, restaurant recommendations, table reservations, and navigation for local life services through natural language. JD is also moving in this direction. Public reports show that JD's AI shopping assistant surpassed 150 million annual active users in 2025, with user penetration exceeding 20%, and drove billions in GMV. Looking at these actions together, it's clear. Platforms are pushing AI from "answering questions" to "handling consumption needs." Brands need to re-understand recommendation In the past, brands mainly relied on several methods for recommendations. First, advertising;

Second, influencer seeding;

Third, platform search and recommendation;

Fourth, offline display;

Fifth, sales guidance and promotions. The common point of these methods is that brands directly or indirectly influence consumers. With AI, there's an extra layer of judgment in the middle. When a user says "Help me buy a case of tissue suitable for the elderly," AI will first understand the need, then filter products. In this process, whether the brand is selected depends on many pieces of information:

  • Whether your product information is clear;
  • Whether your user reviews support this scenario;
  • Whether your price is explainable;
  • Whether your inventory and fulfillment are stable;
  • Whether your brand has enough consumer feedback. So, brand competition will add a new question: Can my product be recommended by AI? This is not just a problem for big brands; small and medium brands also face it. In the past, some brands could sell well by relying on low prices, channels, and shelf interception. If AI starts participating in recommendations, these brands' advantages may be recalculated. Low price is still an advantage, of course. But how cheap, who it's suitable for, whether reviews are good, whether it can be delivered in time, and whether returns and exchanges are troublesome will all enter the judgment. Brands can no longer just talk about concepts. They need to explain the product clearly, the scenarios clearly, and why users repurchase clearly. Retailers need to shift from selling goods to organizing life needs For retailers, the change is also direct. In the past, retailers mainly organized products. Rice, flour, oil, snacks, personal care, household cleaning, beverages, dairy, fresh produce, fruits and vegetables—they displayed them category by category. With AI, users may not search by category. They might say:

"Guests are coming tomorrow, help me prepare some drinks and snacks;

"The child has a spring outing, help me buy some snacks for the road;

"I'm controlling sugar recently, help me choose some suitable breakfast items;

"My parents are coming to stay for a few days, help me replenish some household items." These statements are not standard categories, but they are consumption tasks. If retailers only know how to display categories, they will gradually become passive. They need to connect capabilities like products, inventory, membership, pricing, and delivery so that AI can call upon them. This change is especially important for supermarkets, convenience stores, and home delivery businesses. Because they are closer to consumers' lives and have more complex SKUs. In the past, this was a management difficulty; in the future, it may become an advantage. As long as they can organize products according to life needs, they have the opportunity to capture new traffic. Distributors should not think this is far from them Many distributors may think that AI shopping is a matter for platforms and brands, and not much related to them. This judgment needs to be careful. As long as the front-end consumption entrance changes, the back-end inventory, distribution, pricing, and expenses will follow. If platforms start recommending products based on tasks, brands will require distributors to cooperate with more scenarios. For example, family stock-up packs, instant replenishment packs, office procurement packs, and holiday combo packs. Behind different scenarios, the requirements for inventory, terminal display, pricing systems, and delivery speed are all different. In the past, distributors looked at: whether this month's tasks were completed, which customers paid, and which SKUs were overstocked. In the future, they also need to look at: which products are being called upon by more scenarios, which products have demand in instant retail, which stores have stock but are not being recommended, and which areas have slow delivery affecting orders. AI will not make channel business simpler. It will expose many management problems that were previously hidden, earlier. If your data is inaccurate, inventory unclear, pricing chaotic, and terminal sales weak, in the past you could rely on people to compensate. In the future, once the front end speeds up, the back end will find it harder to keep up. This is what the consumer industry really needs to pay attention to So, the impact of AI on the consumer industry cannot be seen only in whether it can write a Xiaohongshu note or generate a promotional poster. These are certainly useful, but they are not the most core change. The more core change is that the way consumers express their needs is changing. In the past, consumers used a search box to find products. In the future, consumers may hand a sentence to AI to process. This will trigger a chain reaction. The entrance changes, so the shelves change;

The shelves change, so the brand recommendation logic changes;

The recommendation logic changes, so product information, review data, inventory fulfillment, and pricing systems will all be re-examined;

Further, sales organizations and distributor management will also be affected. This is why I say AI is reorganizing consumer decision-making. Not all categories will change immediately, and not all consumers will change their habits right away. The consumer industry is large and slow, and many old logics will continue to exist for a long time. But once there is an additional entrance, competition will gradually change. Whoever can enter consumers' life needs earlier, whose products can be clearly explained, whose goods can be found in time, and whose fulfillment is more stable, will be more likely to be included in the next round of choices. For brands, retailers, platforms, and distributors, this is just the beginning. But it is no longer just a technical issue. It is becoming a business issue.