Doubao launches water and gas bill payment in Shanghai as large models compete for lifestyle services entry points

The "battle" of AI large-model competition has already "burned" into many corners of daily life. Recently, a Shanghai user discovered while using Doubao that saying "help me pay this month's water and gas bills" allowed the system to recognize the request, call the service, and complete the payment. This means that large-model daily-life bill payment services are now available in the Shanghai area. An apparently inconspicuous bill-payment entry point has torn open a new front in large-model competition: AI is transforming from a "chat tool" into an "entry point for daily-life services." Some experts believe that the reason giants are eyeing this so covetously is not only the enormous potential of daily-life service consumption, but also that once "handling affairs with a single sentence" is achieved, the original marketing logic will undergo drastic change. Therefore, companies must seize the entry point in advance and rebuild their ecosystems. Giant competition, with the "battle" spreading to high-frequency scenarios During this year's Mid-Autumn and National Day holiday, a consumption report released by Douyin Life Services showed that the AI entry point became a new holiday consumption scenario for the first time, with order volume for "Doubao hotel booking" rising more than 56% month on month. Behind this is Doubao's recent completion of integration with Douyin Life Services: users can make requests in natural language within Doubao and receive recommendations for stores and group-buy packages, and can directly complete payment, view orders, and retrieve redemption codes. Partners include Flight Manager, High-Speed Rail Manager, Caocao Mobility, and Baidu Maps. When a user says, "Help me book a flight from Beijing to Shanghai tomorrow," Doubao automatically recognizes the intent and calls the corresponding service. A relevant person confirmed that Doubao is gradually building capabilities for handling affairs in daily-life scenarios such as travel services and utility bill payments. Taking payment of local daily-life bills such as water, electricity, and gas as an example, users can use this function after binding their Douyin account. However, the availability of payment items currently differs among cities: Shanghai has covered two payment items, water and gas bills, while Beijing has opened a channel for water bill payment, and services are still being connected gradually. In fact, the struggle among large models over daily-life service scenarios showed signs as early as the beginning of the year. In January this year, Qwen announced its integration with Fliggy, enabling "book flights and hotels with a single sentence," and subsequently fully integrated with Alibaba ecosystem scenarios such as Taobao Flash Sale, Amap, Taobao, and Alipay; in February, it further integrated with shopping and performance platforms such as Tmall Supermarket, Freshippo, and Damai, forming a relatively large ecosystem. At that time, Wu Jia, president of Qwen's consumer business group, demonstrated "ordering with a single sentence" to reporters. He issued an instruction through Qwen: "Help me order 40 cups of Chagee's Boya Juexian," and Qwen immediately called Taobao Flash Sale to place the order and completed payment with one click through the built-in "Alipay AI Pay," without any redirects. The fierce competition among internet companies in the daily-life services field back then is now being replayed in the large-model era, and no one will give up easily. As a giant in local life services, Meituan has chosen deep cooperation with Tencent. On the earnings call for the first quarter of 2026, Meituan founder Wang Xing revealed that Meituan's agent "Xiaomei" will cooperate deeply with Tencent Yuanbao. When users submit local service requests in Tencent Yuanbao, the system will seamlessly connect to Meituan's local service ecosystem, including food delivery ordering and delivery. "In the future, service agents are becoming increasingly important." Wang Xing revealed that Meituan's AI assistant "Xiaotuan" has been placed in the core position of the bottom navigation bar. Logic reset, multi-terminal entry points will become a trend Unlike the mobile internet era, when mobile phones were "dominant," reporters noted that the "brains" of large models will appear in more terminal forms. This year's Inclusion Conference showcased a broader service landscape: AI is beginning to step out of mobile phones and take over people's "life outside 8 hours." For example, put on AI glasses and say, "Help me check nearby coupons," and discount dining vouchers are pushed in front of your eyes, with payment completed by voice; sit in a Li Auto vehicle and say to the in-car assistant, "Help me pay the parking fee" or "Help me buy a movie ticket," and the barrier gate lifts, movie seats are selected, and gas station navigation can all be completed with one click in the car. At the same time, large models are also breaking through in terms of temporal continuity, moving from instant response to continuous execution on behalf of users. Take the newly launched "Coffee Alarm" feature on Alipay as an example: users only need to say, "Every day at 10 a.m., buy a Starbucks latte," and the large model will remember this and automatically start work every day to execute the task, until the final step of waiting for payment confirmation. These seemingly scattered scenarios have, to some extent, reset the underlying logic of daily-life services. AI large models are connecting mobile phones, car systems, and smart glasses through cross-terminal interconnection protocols, allowing services to follow "intent" rather than applications. From the perspective of the competitive landscape among giants, this means the battle for entry points is spreading from mobile phones to multiple terminals. The ByteDance ecosystem continues to expand its territory through Doubao, Alibaba connects its own ecosystem through Qwen, and Alipay's "Abao" is trying to use cross-terminal protocols as a "connector," allowing different agents to complete handshakes at the payment and fulfillment layers. According to the "2026 China Mobile Internet First-Half Report" released by a market research institution, concentration in the AI-native app market continues to increase. As of June, the monthly active user scales of Doubao, Qwen, and DeepSeek reached 382 million, 167 million, and 130 million, respectively, firmly ranking in the first tier of domestic consumer AI applications. Changing the rules, ecosystem and "breaking down walls" are more critical Why are giants vigorously competing for daily-life service entry points? In the view of internet analyst Zhang Shule, local life services are the field most convenient for reaching users' rigid demand, and also the experience entry point for large models to shift from a "chat companion" identity to a "butler" identity. Moreover, judging from the "doing work" capabilities of large models, this also fits the needs of various platforms' local life intelligent service layouts, from easy to difficult. Han Xinyi, CEO of Ant Group, believes that agent commerce will see an explosion in the next 6 to 12 months. "For the past 20 years, people learned to use the language of machines to express their needs; in the next 20 years, models will learn to use human language to understand needs." It is worth noting that when large models can "complete transactions with a single sentence," the traffic allocation rules of traditional application entry points will also change. For large models, ecosystem building has become even more critical. In August, the Qwen Open Platform launched, opening access to three types of terminal services—mobile phones, PCs, and AI glasses—to ecosystem partners and developers. It has currently integrated high-frequency daily-life service applications such as SF Express, Tiandao Home Services, VariFlight, Ziroom Rental, and Hello Car Rental. As of September, the service coverage of this open platform has expanded to more than 20 fields, with more than 1,000 partners applying to join. For example, the SF Express agent can already remember logistics relationships such as users' shipping addresses and directly complete express delivery requests such as "Help me send a document to a client in Shanghai." However, how to move from ecosystem building to application and form a complete task chain—especially how to "break down walls" among the Alibaba, Tencent, and ByteDance ecosystems—remains the biggest challenge facing large models. In the long run, Lu Ming, a member of the National Committee of the Chinese People's Political Consultative Conference and a distinguished professor at Antai College of Economics and Management, Shanghai Jiao Tong University, is optimistic. He believes that "AI + consumption" is not only the best "test field" for technology implementation, but also the most tangible economic calculation. "If AI agents can cross different scenarios, effectively bridge the 'breakpoints' between service scenarios, and use voice and other methods to close the 'digital divide' for the elderly and lower-tier markets, they can greatly lower decision-making thresholds and proactively create and stimulate new consumer demand."