📄 中文摘要
最近，万事达卡和谷歌开源了一个名为“可验证意图”的加密框架，旨在证明 AI 代理的行为符合人类的授权。同时，Ramp 推出了“代理卡”，为 AI 代理提供了独立的企业信用额度，并设定了消费限制和商户限制。这些不同公司的不同方法揭示了一个共同的认识：自主 AI 商务的缺失关键不是能力，而是信任。因此，构建验证机制的公司正在为未来的代理支付基础设施铺路。
未建立的信任层：为何 AI 代理在消费前需要验证
📄 English Summary
The Trust Layer Nobody Built: Why AI Agents Need Verification Before They Can Spend
This week, two significant developments highlighted a gap in agentic infrastructure. Mastercard and Google open-sourced a cryptographic framework called 'Verifiable Intent', which ensures that an AI agent acts according to human authorization. Meanwhile, Ramp launched 'Agent Cards', providing AI agents with their own corporate credit lines, complete with spending limits and merchant restrictions. These different approaches from various companies reveal a shared realization: the missing piece for autonomous AI commerce is not capability, but trust. Companies focusing on verification are strategically positioning themselves for the future of agentic payments.