AI 模型和平台
Uber 称 AI Agents在万圣节周期间只订了四条车厢
在 Google 上的首选来源中添加 Unite.AI
TechCrunch在2026年11月4日报道称，Uber只记录了4次AI-agent订票的旅程，大约500次agent放置的Uber在万圣节周内吃到订单，这是10月最繁忙的一周，而CEO德拉·霍斯罗沙希仍然将外部助理视为潜在的需求来源。
首席财务官Balaji Krishnamurthy告诉分析人员，Uber的第三季度收入电话称，该公司在该周处理了大约3.2亿次旅行和订单，跨越Uber和Uber Eats。 显示的当前交易量少之又少
Khosrowshahi将agent的牵引力描述为非常早且有限，但对成长仍然充满希望，他把第三方agents看作是达到Uber平台的另一种方式，也是生态系统需求的潜在净正值。
Khosrowshahi估计市场上约有8名主要的AI助手，预计人数将超过10人。 每个AI系统公司都在与Uber交谈，指出在大量agent驱动的交易量到来之前，对发行有相当的兴趣。
Uber的第三季度结果提供了更广泛的操作环境。 该公司记录了41亿次旅行，衡量完成的流动旅行和交货订单，年增长率为16%。 月活跃平台消费者，完成搭乘或接受送货的独特月用户的季度平均值达到2.15亿，上升14%.
订书速度、发现和完成
Uber的预言将中位预订时间定在50秒左右，在订购流量中，85%的骑手在离职前积极调整骑行选择，超过三分之一的Uber Eats订单来自顾客在前一年没有订购的一家餐馆。 这些数字共同解释了Uber将AI动力方便与选择和视觉发现相结合的强调。
业务上的区别扩大了这一论点。 agent可以帮助消费者决定去哪里或订购什么；移动仍然需要可用的驱动，实时匹配，定价，皮卡位置和支持。 交货需要商家、信使和消费者，以及存货供应、订单执行和管理交货条件。 Uber认为，无论交易从何处开始，这些业务能力仍将是有价值的。
它与消费者，商人和收入者的直接关系也提供交易历史，个人偏好和实时信息。 这些投入支持Uber努力更快、更方便地预订和订购，让自己的应用程序发挥一种作用，而不仅仅是为外部助理提供出货按钮。
Uber自己的助理和外部-Agent术语
Uber 称早期使用Cart Assistance是令人鼓舞的，客户将其用于更复杂的杂货购物和篮子，其大小大约是普通的Grocery & Retail 订单的两倍。 它的Ask Eats β是一个为特定请求设计的语音辅助助手，比如在15美元以下找到一个可以在20分钟内到达的健康午餐。 Ask Eats将消费者的喜好，餐厅的选择，地点和现场交付条件结合起来来产生建议。
Uber将继续在自己的应用程序内投资AI体验，同时有选择地与能够带来增量需求的第三方agents合作。 其公布的评价标准是客户经验，可持续经济学，以及维护消费者从包括Uber One在内的直接关系中获得的利益。
兹Uber 其中一人约有5 500万成员，占流动和交付毛额总订票组合的一半以上。 就流动和交付而言，总订货是包括适用税、通行费和手续费在内的乘车和订货的美元总值。 这一成员基础有助于解释为什么Uber将外部-agent分配与维护直接关系福利挂钩，而不是将每个新的订购渠道视为等同的渠道。
小型的万圣节周数使得外部机会得不到证实。 Uber所声明的术语使得下一阶段取决于agents之外是否能够在不影响客户经验，可持续经济学或支持其现有服务的直接关系的情况下添加需求。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
AI Models & Platforms
Uber Says AI Agents Booked Just Four Rides During Halloween Week
Add Unite.AI to your preferred sources on Google
TechCrunch reported on November 4, 2026, that Uber logged just four AI-agent-booked rides and roughly 500 agent-placed Uber Eats orders during Halloween week, its busiest week in October, while CEO Dara Khosrowshahi still saw outside assistants as a potential source of demand.
Chief financial officer Balaji Krishnamurthy told analysts on Uber’s third-quarter earnings call that the company handled around 320 million trips and orders across Uber and Uber Eats that week. The comparison shows how little of Uber’s current transaction volume begins with an AI agent, even during exceptionally heavy platform activity.
Khosrowshahi described agent traction as very early and limited but remained hopeful about growth, viewing third-party agents as another way to reach Uber’s platform and a potential net positive for ecosystem demand.
Khosrowshahi estimated that there are about eight major AI assistants on the market and expects that number to exceed 10. Every AI-system company is talking to Uber, pointing to considerable interest in distribution before significant agent-driven transaction volume has arrived.
Uber’s third-quarter results supply the wider operating context. The company recorded 4.1 billion trips, its measure of completed Mobility rides and Delivery orders, up 16% year over year. Monthly active platform consumers, the quarterly average of unique monthly users who completed a ride or received a delivery, reached 215 million, up 14%.
Booking Speed, Discovery and Fulfillment
Uber’s prepared remarks put the median booking time at about 50 seconds for a ride and about 90 seconds for a habitual Uber Eats reorder. Within those ordering flows, 85% of riders actively adjust their ride option before checkout, and more than one in three Uber Eats orders comes from a restaurant the customer had not ordered from during the previous year. Together, the figures explain Uber’s emphasis on combining AI-powered convenience with choice and visual discovery.
The operational distinction extends that argument. An agent can help a consumer decide where to go or what to order; Mobility still requires available drivers, real-time matching, pricing, pickup locations and support. Delivery requires merchants, couriers and consumers, alongside inventory availability, order fulfillment and management of delivery conditions. Uber believes those operational capabilities will remain valuable regardless of where the transaction begins.
Its direct relationships with consumers, merchants and earners also provide transaction history, personal preferences and real-time information. Those inputs support Uber’s effort to make booking and ordering faster and more convenient, giving its own apps a role beyond simply supplying a checkout button for an external assistant.
Uber’s Own Assistants and External-Agent Terms
Uber said early usage of Cart Assistant has been encouraging, with customers employing it for more complex grocery shopping and baskets roughly twice the size of an average Grocery & Retail order. Its Ask Eats beta is a voice-enabled assistant designed for specific requests, such as finding a healthy lunch under $15 that can arrive within 20 minutes. Ask Eats combines consumer preferences, restaurant selection, location and live delivery conditions to produce recommendations.
Uber will continue investing in AI experiences inside its own apps while selectively working with third-party agents that can bring incremental demand. Its stated evaluation criteria are customer experience, sustainable economics and preserving the benefits consumers receive from direct relationships, including Uber One.
Uber One has approximately 55 million members, who account for more than half of combined Mobility and Delivery Gross Bookings. For Mobility and Delivery, Gross Bookings is the total dollar value of rides and orders, including applicable taxes, tolls and fees. That membership base helps explain why Uber is tying external-agent distribution to the preservation of direct-relationship benefits rather than treating every new ordering channel as equivalent.
The small Halloween-week numbers leave the external opportunity unproven. Uber’s stated terms make the next stage depend on whether outside agents can add demand without compromising customer experience, sustainable economics or the direct relationships that support its existing service.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.