内容来源：https://www.technologyreview.com/2025/10/10/1124313/building-connected-data-ecosystems-for-ai-at-scale/

内容总结：

【本报综述】当前，企业正面临日益复杂的数据生态挑战。随着人工智能技术规模化应用的深入，许多企业的IT系统仍呈现碎片化状态，如同"历经数百年的古城遭遇现代交通拥堵"，数据流动效率低下已成为制约AI效能的关键瓶颈。

研究表明，这一问题具有普遍性。高德纳数据显示，2025年仅有不到半数的企业首席信息官认为其数字化举措达成预期目标。究其根源，IDC在2023年的调研揭示，77%的企业仍受困于数据智能应用难题，直接影响了决策效率。

行业实践表明，构建统一的数据集成平台正成为破局之道。以SAP集成套件为例，该解决方案通过三大核心能力助推企业转型：

1. 医疗科技企业西门子医疗通过构建集成层，在严格遵守医疗数据监管要求的同时，实现了财务流程优化与运营洞察实时化；
2. 奢侈品零售巨头哈罗德百货运用预构建连接器与事件网格架构，将集成流程耗时缩减30%，总拥有成本降低40%；
3. 德国福维克集团通过自动化数据流整合，在五年内将数字化销售占比从1%提升至85%，显著加速订单兑现周期。

专家分析指出，现代集成平台正在与AI技术形成良性循环：集成平台为AI提供高质量数据基础，而AI技术反过来通过智能流程设计、异常监测与自修复功能增强集成效能。这种协同效应在零售业个性化服务、医疗诊断影像、供应链优化等领域已显现价值。

值得注意的是，IDC最新报告显示近半数企业仍在使用三种以上集成工具，25%的企业甚至超过四种。技术复杂性与人才缺口成为企业整合数据生态的主要障碍。对此，行业专家建议企业应制定前瞻性集成路线图，从被动应对转向体系化建设，为AI时代的持续创新筑牢数据基石。

（本文基于《麻省理工科技评论》洞察团队与SAP联合研究成果编撰）

中文翻译：

与SAP携手合作
构建互联数据生态系统，实现规模化人工智能

现代集成平台正帮助企业精简碎片化的IT环境，为数据管道适应AI驱动的转型做好准备。

企业IT生态系统常常如同无序扩张的大都市——在持续膨胀的数据流量背景下，陈旧基础设施与尖端新技术相互交织，形成多层叠加的复杂环境。正如驾车穿行于经过现代化改造的千年古城，难免遭遇交通堵塞，企业IT系统也频繁面临数据瓶颈。如今的IT架构包含传统大型机、云原生应用、本地部署系统、第三方SaaS工具以及不断扩张的边缘计算生态。数据在这套拼凑式系统中流转时，常常陷入维护成本高昂且易发生拥堵的连接网络——就像刚从高速路驶出，却迎面遇上正在维修的狭窄石桥。

具有前瞻眼光的企业正转向集中式云端集成解决方案。为构建适应AI优先时代的敏捷系统，这些企业开始采用能支持实时数据流、API管理和事件驱动架构的云端集成方案。

无缝连接至关重要
在AI时代，前述的拥堵场景会带来严重隐患。AI模型依赖洁净、一致且丰富的数据；任何延迟或不一致都会迅速拉低输出质量。碎片化的数据流甚至会拖垮最前沿的AI项目。当连接故障发生时，系统将无法以AI驱动流程所需的规模和速度进行通信。

理想与现实的差距
当数据连接存在风险时，即使最具潜力的AI项目也可能无法创造价值。
数据来源：MIT Technology Review Insights根据IDC 2025年数据编纂

数据潜力与现实效果之间的落差，是众多企业IT负责人亟需弥补的痛点。IDC数据显示，2023年有77%的企业表示数据智能是制约决策速度的长期挑战。高德纳咨询公司指出，这或是2025年仅有不足半数CIO认为当前数字化举措达到或超出业务目标的原因之一。

集成赋能AI——AI反哺集成
AI驱动成果的潜力取决于企业全域高速传输洁净数据的能力。与此同时，AI技术本身也正在重塑集成生态。云原生集成平台开始融合AI能力，实现流程设计自动化、异常检测、智能连接推荐，甚至故障数据管道自修复。这就形成了良性循环：集成赋能AI——AI反哺集成。

除技术优势外，现代集成推动的智能自动化还能提升整体运营效率与跨部门协作。业务流程响应更敏捷，数据实现跨部门流通，团队能更快适应市场与客户需求变化。随着集成平台承担更多常规数据处理工作，人力团队可转向更高价值的战略任务。

构建互联互通的AI就绪型企业
集成平台助力统一从本地到边缘的数据流，确保庞大应用生态中的API治理。

配备知识图谱的预构建连接器进一步加速跨系统连接，实时监控功能更在问题影响业务前提供预测洞察与预警。

现实中已涌现通过精妙集成提升企业敏捷性与AI适应性的典型案例。以下三家企业正运用SAP集成套件优化数据流并简化运营：

• 西门子医疗：在数据精确性、时效性与安全性不容妥协的医疗领域，该公司通过集成方案提升医疗服务的可及性与个性化。面对涵盖诊断、影像、治疗等多元业务场景，其集成层不仅简化财务结算与报告等核心流程，还支持灵活规划与实时运营洞察。该系统实现了跨平台数据无缝调取，避免数据重复存储——这在强监管行业尤为关键。

• 哈罗德百货：这家奢侈品零售商运营着支持伦敦旗舰店与增长中的电商业务的混合IT架构，目前在线销售10万款商品，日均通过数字渠道处理200万笔交易。为简化日益复杂的业务网络，哈罗德采用SAP预构建B2B连接器与事件网格架构，在关键业务流程中协调超600条集成流。

实施SAP方案后，哈罗德将集成相关流程耗时缩短30%，总体拥有成本降低40%。更重要的是，公司构建了能随客户期望与零售技术演进而灵活调整的数据应用中枢。

• 福维克：以智能厨电与清洁系统闻名的德国直销企业，近年完成深度数字化转型。2018至2023年间，其数字化销售占比从1%跃升至85%。

福维克依托SAP方案实现CRM、库存管理、支付处理及授权管理等核心系统数据流自动化。升级后的系统消除了人工文书工作，大幅加速订单到现金周期，提升客户数据准确性与一致性。

集成如何铸就竞争优势
通过应用SAP解决方案，哈罗德与福维克已为AI时代的竞争做好铺垫。福维克的数字化转型推动数字销售增长，哈罗德的数据基础设施则随技术与客户期望持续进化。
数据来源：MIT Technology Review Insights根据SAP 2025年数据编纂

这些案例表明，跨行业AI落地皆以连接能力为基础。以医疗领域为例，在行业快速拥抱AI之际，稳健的集成系统是实现影像诊断、预测性护理等场景的先决条件。严苛的法规要求更需确保跨系统数据的精准处理与可追溯性。

零售领域同样如此：从动态定价、个性化推荐到预测性库存管理，所有AI驱动创新都依赖销售、库存、客户及合作伙伴系统间精准快速的数据流，而事件驱动型集成正是其基石。

在福维克这类直面消费者的商业模式中，集成技术带来了前所未有的个性化体验、实时营销与供应链优化。这种能力帮助D2C企业在动态市场中保持竞争力与应变力——这恰逢其时，因为超过70%的消费者如今期待品牌的个性化服务。展望未来，AI（尤其是生成式AI）将在扩展个性化体验方面发挥关键作用，助力品牌以恰如其分的语气、视觉引导与文案提供场景化定制信息。

集成即差异化优势
IDC最新报告显示，近半数企业同时使用三种以上集成工具，25%的企业在环境中部署超四类工具。

尽管多数企业认同整合价值，技术挑战与人才缺口仍是简化进程的阻碍。另一结构性问题在于：三分之一企业在系统实施阶段才考虑集成方案，这限制了从头设计面向未来数据流的机会。

持续创新与长期敏捷性取决于基础设施能否与企业抱负同步进化。现代集成平台正是实现这种适应性的连接纽带。统一集成战略指引了前进方向：通过制定集成路线图，企业可从被动零散的应对转向构建量身定制、可扩展的基础架构，同时满足当前业务需求与AI驱动创新的未来要求。

当代繁荣的城市并非仅靠扩建公路或零散增建环岛来疏导交通——而是通过彻底重构移动出行理念。企业IT领域同样如此：持续创新与长期敏捷性取决于基础设施能否与企业抱负同步进化。现代集成平台正是实现这种适应性的连接纽带。

了解更多请访问MIT Technology Review Insights与SAP合作的"关键业务计划现代集成"内容中心。

本内容由MIT Technology Review定制内容团队Insights制作，未经本刊编辑部门撰写。

本内容全程由人类作者、编辑、分析师及插画师完成调研、设计与撰写，包括问卷编写及数据收集。可能使用的AI工具仅限经过严格人工审核的辅助生产环节。

英文来源：

In partnership withSAP
Building connected data ecosystems for AI at scale
Modern integration platforms are helping enterprises streamline fragmented IT environments and prepare their data pipelines for AI-driven transformation.
Enterprise IT ecosystems are often akin to sprawling metropolises—multi-layered environments where aging infrastructure intersects with sleek new technologies against a backdrop of constantly ballooning traffic.
Similarly to how driving through a centuries-old city that’s been retrofitted for automobiles and skyscrapers can cause gridlock, enterprise IT systems frequently experience data bottlenecks. Today’s IT landscapes encompass legacy mainframes, cloud-native applications, on-premises systems, third-party SaaS tools, and a growing edge ecosystem. Information flowing through this patchwork gets caught in a tangle of connections that are costly to maintain and prone to snarls—sort of like emerging from a high-speed expressway to a narrow, cobblestone bridge that's constantly undergoing repairs.
Forward-looking organizations are now turning to centralized, cloud-based integration solutions.To create more agile systems suited for an AI-first future, forward-looking organizations are now turning to centralized, cloud-based integration solutions that can support everything from real-time data streaming to API management and event-driven architectures.
Seamless connectivity matters
In the AI era, congestion like the scenario described above is a serious liability.
AI models depend on clean, consistent, and enriched data; lags or inconsistencies can quickly degrade outputs. Fragmented data flows can undermine even the most cutting-edge AI initiatives. And when connectivity snafus occur, systems aren't able to communicate at the scale or speed that AI-driven processes demand.
The gap between potential and reality
Even the most promissing AI initiatives can fail to deliver value when data connectivity is at risk.
Source: Compiled by MIT Technology Review Insights, based on data from IDC, 2025.
This gap between data potential and reality is a pain point many enterprise IT leaders struggle to bridge. According to IDC, in 2023, 77% of organizations report that data intelligence is a persistent challenge hampering decision velocity. This may be part of the reason why, according to Gartner, in 2025 less than half of CIOs say their current digital initiatives are meeting or exceeding business outcome targets.
Integration enables AI—and AI, in turn, turbocharges integration.AI’s potential to drive such outcomes hinges on a company's ability to move clean data, at speed, across the entire enterprise. At the same time, AI itself has the potential to reshape the integration landscape. Cloud-native integration platforms are beginning to incorporate AI-powered capabilities that automate flow design, detect anomalies, recommend optimal connections, and even self-heal broken data pipelines. This creates a virtuous cycle: integration enables AI—and AI, in turn, turbocharges integration.
Beyond the technical benefits, intelligent automation facilitated by modern integration stands to improve overall operational efficiency and cross-functional collaboration. Business processes become more responsive, data is accessible across departments, and teams can adapt more quickly to changing market or customer demands. And as integration platforms handle more of the routine data-wrangling work, human teams can shift focus to higher-value priorities.
Building connected, AI-ready enterprises
Integration platforms help unify data streams from on-prem to edge and ensure API governance across sprawling application landscapes.
Pre-built connectors enriched with knowledge graphs further accelerate connectivity across diverse systems, while real-time monitoring provides predictive insights and early warnings before issues impact business operations.
We're already seeing real-world examples of how thoughtful integration is empowering enterprises to become more agile and AI-ready. Here are three companies using SAP Integration Suite to streamline data flows and simplify their operations.

• Siemens Healthineers: In the healthcare sector, where data accuracy, timeliness, and security are non-negotiable, Siemens Healthineers is using integration solutions to make health services more accessible and personalized.
  Siemens Healthineers operates a diverse business landscape spanning diagnostics, medical imaging, and therapy, each with unique data requirements and processes. To enable more autonomous decision-making, the company’s integration layer helps streamline core financial processes, such as closing and reporting, while also supporting flexible planning and instant insights into operations. It also enables seamless data access across systems without the need for data replication, an important consideration in a highly regulated industry. - Harrods: Luxury retailer Harrods operates a complex hybrid IT landscape that supports both its flagship London store and a growing e-commerce business; the company now offers 100,000 products online and processes 2 million transactions per day through digital channels. To modernize and simplify this growing footprint, Harrods leverages SAP's pre-built B2B connectors and Event Mesh architecture to orchestrate more than 600 integration flows across key business processes.
  Since implementing the SAP solutions, Harrods has reduced integration-related process times by 30% and cut total cost of ownership by 40%. More importantly, the company has created a nimble data and application backbone that can adapt as customer expectations — and digital retail technologies — evolve. - Vorwerk: German direct-sales company Vorwerk, known for products like smart kitchen appliances and cleaning systems, has undergone a sweeping digital transformation in recent years. Between 2018 and 2023, the company grew its digital sales from just 1% to 85%.
  Vorwerk relies on SAP solutions to automate data flows across critical systems, including CRM and inventory management, payment processing, and consent management. The updated system has helped eliminate manual paperwork, significantly accelerate order-to-cash cycle times, and improve the accuracy and consistency of customer data.
  How integration can be a competitive advantage
  Using SAP solutions, retailers Harrods and Vorwerk are primed for success in the AI era. Vorwerk's digital Harrods data infrastructureDigital growth
  transformation boosted
  digital salesProcess efficiency
  evolved with technology
  and customer expectations
  Source: Compiled by MIT Technology Review Insights, based on data from SAP, 2025.
  As these examples demonstrate, connectivity is essential groundwork for AI across just about every industry. As the healthcare sector rapidly embraces AI, for instance, robust integration is a prerequisite for use cases like diagnostic imaging and predictive care. Stringent regulatory requirements also demand accurate, transparent data handling and traceability across systems.
  In retail, too, unified, event-driven integration underpins AI-driven innovations ranging from dynamic pricing and personalized product recommendations to predictive inventory management—all of which require fast, accurate data flows across sales, inventory, customer, and partner systems.
  And in direct-to-consumer models like Vorwerk’s, integration enables new levels of personalization, real-time marketing, and optimized supply chains. Such capabilities can help D2C businesses stay competitive and responsive in highly dynamic markets — a necessity as more than 70% of consumers now expect personalized experiences from the brands they buy from. Moving forward, AI (particularly generative AI) will likely play a pivotal role in scaling these personalized experiences and enabling brands to deliver tailored messages with the right tone, visual guides, and copy to meet the moment.
  Integration as a differentiator
  According to a recent IDC report, nearly half of enterprises are juggling three or more integration tools, with 25% using more than four across their environments.
  While many companies see value in consolidating, technical challenges and skill gaps remain barriers to simplification. Another structural issue: One-third of enterprises don’t consider integration until system implementation is already underway—limiting opportunities to design future-ready data flows from the start.
  Sustained innovation and long-term agility depend on whether infrastructure can evolve as quickly as a company's ambitions. Modern integration platforms provide the connective fabric that makes this kind of adaptability possible.A unified integration strategy offers a path forward. An integration roadmap can help companies shift from reactive, piecemeal efforts to a more purpose-built, scalable foundation—one that supports both current business needs and the demands of AI-driven innovation.
  The cities that thrive today aren't the ones that simply manage traffic flow by expanding their highways or adding in sporadic roundabouts—they're the ones that have reimagined mobility entirely. In enterprise IT, the same principle applies: Sustained innovation and long-term agility depend on whether infrastructure can evolve as quickly as a company's ambitions. Modern integration platforms provide the connective fabric that makes this kind of adaptability possible.
  Learn more on the MIT Technology Review Insights and SAP Modern integration for business-critical initiatives content hub.
  This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.
  This content was researched, designed, and written entirely by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.