企业 AI 采用正在成长，目前已有 88 percent 的组织至少在一项功能中使用 AI，但许多企业只看到中等幅度的收益，或在少数情况下，根本没有任何可衡量的价值。2025 年全球受访的四分之三 CEO 表示，如果这种情况不改变，他们担心自己可能会在两年内失去工作。研究显示，根据 AI 的能力从根本上重新设计工作流程的组织，更有可能充分享受其潜力，而从把 AI 当作外加技术转向 AI-native，被形容为思维上的剧烈转变。

Benedict Evans 说，对于生成式 AI，人们应该预设它会带来与网际网路和行动装置同等规模的变化，而 AI 时代最成功的公司将会发明全新的 AI-native 产品、服务与营收策略。他说，领导者应该思考当某件昂贵的事物变得明显更便宜或免费时会发生什么，考虑电脑过去无法处理哪些类型的问题，并且不要一开始就把反直觉的想法一概否定。他也说，AI 未来可行性的物理限制仍然未知，战略性探究必须持续演进。

Michael Bradshaw 说，一家公司的技术策略就是它的商业策略，但大多数企业 IT 建构于不同的时代，而且资料与应用程式往往在整个企业内彼此孤立。他说，AI-native 的成功需要能横向运行于整个业务的统一资料平台，以及模组化、可组合的架构，而不是孤岛式、受供应商绑定的系统，而且 agentic AI 可以跨系统游走，并即时协助设计工作流程。Cassie Kozyrkov 说，领导者常把 AI 当成仿佛具有自主性来看待，但它更关乎决策、提示词、政策、护栏，以及对结果的定期审查；她还说，领导力是生成式 AI 系统中的一项技术性输入。

Corporate AI adoption is growing, with 88 percent of organizations now using AI in at least one function, but many businesses are realizing only moderate gains or, in a small proportion of cases, no measurable value at all. Three out of four CEOs polled worldwide in 2025 said they feared they could lose their jobs within two years if that picture does not change. Research shows that organizations that fundamentally redesign workflows in light of AI’s capabilities are more likely to enjoy its full potential, and the move from AI as a bolt-on technology to AI-native is described as a dramatic shift in mindset.

Benedict Evans says that with generative AI people should presume the same scale of change that came with the internet and mobile, and that the most successful companies of the AI era will invent entirely new kinds of AI-native products, services, and revenue strategies. He says leaders should ask what happens when something expensive becomes significantly cheaper or free, consider what kinds of questions computers could not previously address, and not dismiss counterintuitive ideas out of hand. He also says the physical constraints of what will be possible with AI are still unknown and that strategic enquiry will need to keep evolving.

Michael Bradshaw says a company’s technology strategy is its business strategy, but most enterprise IT was built in a different era and data and applications are often siloed across the enterprise. He says AI-native success demands unified data platforms that run horizontally across the business and modular, composable architectures rather than siloed, vendor-bound systems, and that agentic AI can roam across systems and help design workflows in real time. Cassie Kozyrkov says leaders often approach AI as if it were autonomous, but it is much more about decisions, prompts, policies, guardrails, and regular review of outcomes, and she says leadership is a technical input into generative AI systems.