内容来源：https://news.microsoft.com/source/emea/features/building-qiddiya-city-how-copilot-helps-abdulrahman-alali-navigate-a-project-of-unprecedented-scale/

内容总结：

微软Copilot助力沙特奇迪亚超级新城建设，AI成巨型项目管理“智慧大脑”

在沙特阿拉伯首都利雅得西南约45公里处，一项规模空前的超级工程——奇迪亚新城正在广袤的沙漠中迅速崛起。这座规划面积达360平方公里（相当于三个巴黎市区）的未来之城，旨在打造集娱乐、体育与文化于一体的巨型目的地，包含主题公园、购物中心、体育场馆、F1赛道、博物馆等设施，未来还将容纳50万居民。

面对如此浩大的工程，项目管理复杂度堪称“噩梦”。奇迪亚投资公司（QIC）首席技术官阿卜杜勒拉赫曼·阿拉里坦言，项目涉及700家公司、2.2万名工人，不同团队使用20多个互不兼容的系统，甚至对同一栋建筑的命名规则都各不相同。

“我们正在利用微软Copilot来总结差异，并为我们提供统一设计标准的最佳方案，”阿拉里表示。通过微软Power BI仪表板与Copilot的深度整合，项目人员只需输入自然语言指令，就能从海量数据中即时提取关键信息，例如快速定位“逾期超过60天且工程师未备注”的发票。

据统计，Copilot已为QIC带来显著效能提升：每月自动生成约25万封邮件与聊天摘要，四个月内累计汇总超5万场会议纪要、生成1.3万份基于企业数据的文档。更重要的是，AI正深入业务核心——从分析主题公园游客体验数据，到研判客户整体满意度，Copilot已成为挖掘数据价值的“超级助手”。

阿拉里强调：“技术能为善用它的企业带来巨大优势。关键在于以正确方式规划AI落地——只要在实施前投入足够时间进行设计与规划，就能获得超乎预期的回报。”随着奇迪亚新城这座“未来之城”轮廓日渐清晰，其背后的AI管理范式或将为全球超大型复杂工程提供全新解决方案。

中文翻译：

建设奇迪亚城：Copilot如何助力阿卜杜勒拉赫曼·阿拉尼驾驭规模空前的超级工程

在距离沙特阿拉伯首都利雅得约45公里处，一座旨在打造娱乐、体育与文化目的地的特大型定制化城市正拔地而起。奇迪亚城占地360平方公里（约140平方英里），面积相当于三个巴黎。这里将建设主题公园、购物中心、体育场馆、一级方程式赛道、博物馆等设施，未来还将在20个社区容纳50万居民。这项浩大工程涉及700家企业和2.2万名工人，而Microsoft 365 Copilot正帮助管理者统筹这座超级城市的千头万绪。

"我们正在同步建设体育场馆、赛道、酒店等多种设施，"奇迪亚投资公司（QIC）首席技术官阿卜杜勒拉赫曼·阿拉尼表示。该公司负责统筹城市建设。为掌握所有承包商的工作进展——从项目计划到已付、待付或暂缓的发票——各类信息经由不同工具汇总至Microsoft Power BI创建的仪表板。通过Copilot，奇迪亚投资公司的员工可以直接提问，从海量工程数据中快速提取关键信息。

管理超级工程的工作流程

阿拉尼进一步解释道，约200人的团队负责制定规划，另有约100人的团队负责项目跟踪与执行，两组人员使用的20个系统采用不同编程语言且互不兼容。更复杂的是，购物中心或游乐园等单体项目可能包含数十栋建筑，每栋建筑都配有20至30位字符的识别码。

"我们发现两个团队对资产（建筑、主干道、街巷）的命名规则完全不同，"他说，在不同系统中匹配资产名称"简直是场噩梦"。"因此我们正运用Copilot分析差异，为统一城市设计标准提供最优方案。这个应用场景是我们上个月刚构思的，目前已在实施过程中。"

跨团队推广Copilot应用

通过Power BI仪表板，奇迪亚投资公司的房地产开发和施工团队使用Copilot已近一年，其他部门则在过去几个月开始采用，主要用于提升工作效率，如汇总邮件、撰写报告等。阿拉尼透露，Copilot每月为该公司自动生成约25万封电子邮件和聊天记录，四个月内累计总结超5万场会议，并基于企业数据创建了1.3万余份文档。

Copilot带来的不仅是效率提升。"由于我们拥有海量建设项目，且每个资产（无论是主题公园、博物馆还是体育场）都深度融合了科技元素，我们正运用Copilot进行研究，探索提升客户体验的最佳方案，比如分析主题公园游客的实际行为模式，"阿拉尼说，"同时通过Copilot解析客户整体情绪相关的数据库。"

在阿拉尼看来，Copilot在涉及海量数据的工程环节都能发挥重要作用——对于超级工程而言，几乎每个环节的信息都多到人力无法处理。"可以想象我们每天收到的发票和付款凭证数量，那是成千上万的单据。"Copilot能够"通过简单指令处理海量数据，例如'找出逾期60天以上且工程师未备注的发票'。有些发票可能因特定问题被故意延迟支付，这是仪表板无法识别的。"

但Copilot可以检索付款凭证和发票的备注信息，比如筛选出600张逾期60天以上的发票，并指出其中10%存在待处理的备注或问题。"这提供了极大的帮助，"阿拉尼强调。

"关键在于以正确方式应用Copilot乃至整个人工智能技术。如果在规划与设计实施阶段投入足够精力，就能最大化发挥其价值，"阿拉尼总结道，"我坚信，科技将成为善用者企业的决定性优势。"

英文来源：

Building Qiddiya City: How Copilot helps Abdulrahman AlAli navigate a project of unprecedented scale
About 45 kilometers from Riyadh, the capital of Saudi Arabia, construction is rapidly advancing on the largest purpose-built city intended to be an entertainment, sports and cultural destination.
Qiddiya City will cover 360 square kilometers (140 square miles), making it three times the size of Paris. With theme parks, shopping, sports venues, a Formula 1 racetrack, museums and more, it will also eventually house 500,000 residents in 20 neighborhoods. The massive project involves 700 companies and 22,000 workers. Microsoft 365 Copilot is helping to keep track of Qiddiya City’s many moving parts.
“We’re building multiple assets, like stadiums, sports tracks, hotels,” says Abdulrahman AlAli, chief technology officer at Qiddiya Investment Company (QIC), which oversees the city’s construction.
To keep track of what all the contractors are doing – from project plan, to paid, pending or on-hold invoices – the information goes through different tools and then into a dashboard created by Microsoft Power BI. Copilot then allows QIC employees to ask questions and quickly dig out relevant information from the megaproject’s mega-pile of data.
Managing a megaproject’s workflow
In addition, AlAli explains, a team of about 200 people is building the plans, while a different team of about 100 people is tracking and executing the project, using 20 different systems among them that use different programming languages – and don’t understand each other. On top of that, assets such as shopping centers or amusement parks may each encompass dozens of buildings, and each building has an ID of 20 to 30 characters.
“We found that the naming of assets – the buildings, the roads, the streets – is different between the two teams,” he says. “It’s a nightmare” to try to match up the asset names across the different systems. “So, we’re using Copilot to summarize the differences and give us the best options to unify the design standards for our city,” he said.
“That’s one use case that we came up with just last month, and we are executing this as we speak,” he added.
Scaling Copilot across teams
QIC’s real estate development and construction teams have been using Copilot for about a year through the Power BI dashboard, and the rest of the workforce has adopted it only in the past few months, mostly for productivity, such as summarizing emails or writing reports. Copilot has autogenerated about 250,000 email messages and chat interactions a month for QIC, summarized more than 50,000 meetings and created more than 13,000 documents based on corporate data in four months, AlAli says.
Copilot provides more than efficiency to QIC. “Because we have massive, massive programs, and technology is embedded in every asset we build, whether it’s a theme park or a museum or a stadium, we’re using Copilot for research, how to find the best ways to enhance customer experience, to understand the actual usage of attendees, for example, for theme parks,” AlAli says. “And we’re using Copilot to interrogate databases related to customer sentiment overall.”
AlAli sees a big role for Copilot in any aspect of the project that involves mountains of data – and with a megaproject, almost every aspect has too much information to humanly sift through. “You can imagine the number of invoices and payment certificates we get every day. Thousands of payment certificates,” he says. Copilot provides the ability to “interrogate all this massive data with a prompt: ‘Give me invoices that are more than 60 days late and there are no comments from the engineers.’ Certain invoices will have a snag listed on them that has been delayed for payment on purpose and the dashboard will not know that.”
But Copilot can look for comments on the payment certificates and invoices and say that, for example, 600 invoices are more than 60 days late, and that out of those, 10 percent have comments, or snags, that need to be addressed. “That’s very, very helpful,” AlAli says.
“It’s about the right way to adopt Copilot and AI in general. If you spend time in planning and designing the implementation, you would get the best out of it,” AlAli says. “I’m a big believer that technology is an unfair advantage for companies that use it.”