rsssoftware

从Token消耗到日活智能体数（DAA）产出：智能经济时代企业AI价值的新度量衡

Lianfeng Wu — Wed, 29 Jul 2026 04:00:00 GMT

本报告围绕DAA如何成为智能经济时代企业AI价值新度量衡展开。第一部分阐述DAA为什么出现，以及它如何标志AI竞争从模型能力转向任务执行力。第二部分梳理从DAU、Token到DAA的度量体系迁移，说明Token是成本语言，DAA是智能生产力活跃语言，业务结果才是财务价值语言。第三部分从CEO、CIO、CFO、COO、CMO、CHRO六类CxO视角，分析DAA对企业战略、架构、成本、运营、增长和组织重构的意义。第四部分提出DAA六层指标树，避免企业把DAA做成数量游戏。第五部分从金融、制造、零售与消费品、政务公共服务、能源与交通五大行业拆解DAA落地场景。第六部分提出DAA运营底座，包括模型、数据、工具、编排、身份权限、评测监控、运营商业七层架构。第七部分给出0—6个月、6—18个月、18—36个月三阶段落地路线图。第八部分分析百度在DAA方向的产品、技术、生态与应用案例，为行业用户选择智能体平台提供评估参考。

IDC中国副总裁兼首席分析师武连峰表示：“AI超级周期下，企业不能只关注模型是否先进、Token消耗多少，更要关注AI是否真正进入业务流程、完成任务并创造结果。DAA的价值在于把智能体从‘技术工具’转化为可管理、可运营、可衡量的数字劳动力，帮助企业建立面向智能经济时代的新型AI价值度量体系。”

IDC Survey: The Martech Stack in Transition — AI-Native Vendors’ Race to Catch Budget, Buyers, and the Baton

Roger Beharry Lall — Tue, 28 Jul 2026 04:00:00 GMT

This IDC Survey examines how U.S. organizations are transitioning from traditional martech stacks toward AI-native marketing platforms. Drawing from an online survey of over 100 senior marketing, advertising, and IT/data leaders across small, midmarket, and large organizations, the survey focused exclusively on those already using, piloting, planning, or evaluating AI-native marketing applications. The research maps current stack composition against the trajectory organizations expect over the next 12 months.

The study identifies the drivers pulling organizations toward AI-native tools and the barriers holding broader adoption back, and it quantifies how far marketing budgets are shifting from traditional to AI-native vendors, now and a year from now. It also maps where AI-native adoption is strongest and weakest across core marketing functions, examines the rise of internally built AI tools alongside vendor-supplied ones, and looks at preferences for orchestration and workflow architecture as the stack evolves.

For technology vendors, services firms, and marketing leaders, the findings provide a benchmark for where the midmarket stands on the path to AI-native martech, and they pinpoint where budget is migrating, which use cases each segment is adopting first, what continues to hold deployment back, and which layer of the stack is best positioned to take on the orchestration role.

IDC TechScape: Worldwide Frictionless Experience-Enabling Innovation in Air Travel, 2026

Dorothy Creamer — Tue, 28 Jul 2026 04:00:00 GMT

This IDC TechScape maps 24 technologies shaping the frictionless airline traveler experience in 2026, organized across transformational, incremental, and opportunistic adoption phases. The study covers the full passenger journey, from booking and pre-trip planning through security, in-flight experience, luggage handling, and post-trip engagement, as well as the crew- and operations-facing technologies that directly determine whether travelers encounter a friction-free journey or a frustrating one.

“Passenger and traveler demands are at an all-time high and the opportunities for disruption and frustration lurk at every digital touch point. The airlines and travel providers that are working to support travelers with proactive and predictive engagement and operational strategies will win with contextualized and frictionless experiences every time,” says Dorothy Creamer, senior research manager, Hospitality and Travel Digital Strategies, IDC. “The constraint separating leaders from laggards in air travel is no longer which vendor relationships they hold; it is whether their data estate is mature enough to let these technologies work together and deliver on their promise.”

IDC's Worldwide: Dedicated Cloud Infrastructure and Services Taxonomy, 2026

Tue, 28 Jul 2026 04:00:00 GMT

This IDC study provides a detailed taxonomy for the dedicated cloud infrastructure and services market. The document covers the definitions of various building blocks and provides a top-level overview of dedicated cloud segments. IDC is retiring the term private cloud and replacing it with the term dedicated cloud, but the definition remains the same: environments in which IT resources are dedicated to a single organization and managed and delivered to end users utilizing cloud attributes. The growing availability of commercial and open source cloud platforms enables the increasing adoption of dedicated clouds by enterprises, while the proliferation of infrastructure management platforms for managing resources across multiple environments drives the expanding growth of hybrid IT environments.

"Dedicated cloud infrastructure as a service has become the defining ground for the enterprises seeking operational control over data placement, compliance posture, or workload governance," said Jasdeep Singh, research manager for IDC's Cloud and Infrastructure Services. "As OEM vendors, cloud service providers, and managed service providers compete in this market, the conversation has shifted from who can build the cloud to who can govern it, on the customer's terms, at the customer's choice of location, and in compliance with their regulatory and AI risk realities."

Market Forecast: Worldwide Enterprise Resource Management Applications Forecast, 2026–2030 — AI Creating Value

Mickey North Rizza — Tue, 28 Jul 2026 04:00:00 GMT

This IDC presentation provides the five-year forecast for the worldwide enterprise resource management applications market and delivers insights into the drivers and inhibitors of growth.

“AI is the accelerant creating value for the enterprise resource management market. Embedded AI in tasks, workflows, and processes is driving innovation and enabling the future with AI assistants and AI agents. Organizations are in the midst of shifting to Agents as Apps, dynamically reshaping the market and creating new software operating models. This shift increases the organization’s dependency on AI-powered enterprise resource management software as it moves to an agentic market.” — Mickey North Rizza, group vice president, Enterprise Software at IDC

Market Share: Worldwide Enterprise Performance Management Shares, 2025

Megha Kumar — Tue, 28 Jul 2026 04:00:00 GMT

This IDC presentation examines the 2025 market shares of the leading vendors in the enterprise performance management (EPM) software market.

"The EPM market continues to grow to meet the need for decision velocity, consistent reporting, and adhering to compliance requirements. As vendors introduce AI capabilities, especially AI agents, expect an increased focus on data readiness, explainability, and governance in evaluations moving forward," says Megha Kumar, research vice president, Analytics and AI, IDC.

Medidata NEXT London 2026: AI Moves from Pilot Purgatory to Platform Scale

Silvia Piai — Tue, 28 Jul 2026 04:00:00 GMT

Medidata Solutions’ (Medidata) NEXT London 2026 marked the shift from AI experimentation to enterprisewide agentic deployment, anchored by the commercial launch of Medidata Plus and the full rollout of Dot. Key innovations include AI-driven simulation workspaces for protocol and enrollment planning, a proprietary Patient Burden Index, GenAI-powered site selection, and an expanded patient engagement platform that integrates continuous sensor data. Customer evidence, including a 37% enrollment uplift in atopic dermatitis trials and up to 70% reduction in budget build times, confirmed operational impact. Medidata also presented its “orchestrated intelligence” vision, extending clinical research intelligence across the life sciences value chain through partnerships with Anthropic, NVIDIA, and Mistral.

Clinical organizations should prioritize data standardization, adopt simulation and site selection tools early in protocol design, and evaluate contract research organization (CRO) partners on AI readiness. Medical device companies navigating the evolving regulatory environment in 2026 should treat early platform engagement as a strategic opportunity.

“Medidata NEXT London 2026 showed an industry moving from AI ambition to AI accountability. The evidence is real. Enrollment gains, faster builds, smarter site selection, and platform models such as Medidata’s offer one indication of how these capabilities can move from pilots to enterprise scale. For medical device companies, the moment to start engaging is now while both the underlying platforms and regulatory landscape are still being defined,” says Silvia Piai, research director, IDC.

中国人工智能基础数据服务市场研究报告，2025

Yanxia Lu, Leo Li — Tue, 28 Jul 2026 04:00:00 GMT

2025年中国人工智能基础数据服务市场规模达62.62亿元人民币，预计在2025至2030年间将以19.6%的年复合增长率持续扩张。本报告系统梳理了中国AI基础数据服务市场的规模预测、竞争格局、技术趋势与政策环境，并对头部厂商进行了深度分析。从市场趋势来看，AI基础数据服务正经历五大关键转变：从泛多模态到生产行为信息聚焦，从纯人工标注转向人机协同混合模式，从大规模知识采集转向复杂任务数据构建，从通用标注向医生、律师等高价值专家数据延伸，以及从答案标注向推理过程与评测体系升级。

IDC中国研究总监卢言霞表示：”AI基础数据服务市场正从’劳动密集型标注’向’智力密集型数据工程’跃迁，数据服务商必须加速技术升级与人才结构转型，将合成数据作为第二增长曲线，并深度绑定垂直行业头部客户，从’数据供应商’向’模型质量裁判’跨越。”

Driving Operational Excellence with Digital Twins in Manufacturing

Sarah Lee, Carlos Gonzalez, Gunjan Bassi — Mon, 27 Jul 2026 04:00:00 GMT

This IDC Perspective examines how digital twins are evolving from visualization tools into the simulation and orchestration backbone of physical AI on the shop floor. It distinguishes between the simulation role, which models processes and assets to support engineering and AI training, and the orchestration role, which coordinates physical systems, AI agents, and workers in real time. It examines divergent adoption patterns across process and discrete manufacturing and profiles the vendor landscape. It provides guidance for operations and plant leaders on sequencing investments, building data foundations, and pairing twin programs with workforce enablement.

"Digital twins are evolving beyond visualization to become the operational backbone of AI-enabled manufacturing. As manufacturers move toward autonomous operations, digital twins provide the simulation environment to train AI and the orchestration layer to coordinate assets, systems, and people in real time," says Sarah Lee, senior research director, Manufacturing IT Strategies at IDC.

"A digital twin may be technically secure yet still become operationally unreliable if its data, models, and assumptions drift from reality. Trust requires both protection and continuous validation," says Gunjan Bassi, research manager, Worldwide Industrial Internet of Things (IIoT) and Intelligence Strategies at IDC.

"A digital twin is only as good as the data beneath it. Without a unified industrial data layer that mirrors real-world plant conditions, even the most sophisticated twin serves only as a virtual mirror. For real digital twins to evolve in the world of physical AI, real-time data analysis and automation must be in sync with digital twin simulation," says Carlos Gonzalez, research manager, Event-Driven Automation and Analytics at IDC.

Emerging Agentic Automation Patterns

Neil Ward-Dutton — Mon, 27 Jul 2026 04:00:00 GMT

This IDC Perspective examines how AI agents are being applied to business operations, beyond early adoption in marketing and software development. Using IDC's Business Automation Blueprint, the document identifies four emerging patterns — agents as task actors, knowledge assistants, orchestrators, and designers/improvers — and explains where each fits alongside established deterministic automation technologies such as workflow tools, RPA, and integration platforms. It closes with practical guidance for technology buyers evaluating where and how to deploy agents within existing operations.

"As AI agents begin to be used within broader business automation initiatives, it's vital to understand the different roles that AI agents can play, and how those roles differ," said Neil Ward-Dutton, research VP, Agentic Automation and AI Technologies at IDC.