Introduced at Money20/20 Asia, the StableX Know Your Agent Framework establishes how AI agents are identified, authorised, monitored and held accountable to provide financial services in payments, compliance, and wealth management — authored in Singapore, designed for the world.
BANGKOK, THAILAND, 21 APRIL 2026 – MetaComp Pte. Ltd. (MetaComp), today launched the StableX Know Your Agent (KYA) Framework – a governance framework for AI agents operating in regulated financial services in payments, compliance, and wealth¹ workflows authored by a licensed financial institution and believed to be the first of its kind globally. MetaComp is Asia’s pioneer in unified Web2.5 digital financial solutions bridging fiat and stablecoin capabilities across payments, treasury, and wealth¹ management through a group-level platform. The KYA framework is open for adoption by financial institutions, regulators, and network partners.
The announcement was made at Money20/20 Asia in Bangkok, Thailand, alongside the expansion of MetaComp’s AgentX agentic financial services Skill ecosystem, the first such ecosystem from a regulated financial institution, which will be available across Claude, Claude Code, OpenClaw, and other compatible AI platforms from 21 April 2026 at www.metacomp.ai.
Ms Tin Pei Ling, Co-President, MetaComp said:
“AI agents are already operating in financial services — initiating payments, making compliance decisions, managing portfolios. And yet there is no agreed standard for who those agents are, what they are permitted to do, or who is accountable when they act outside their mandate. KYA is our active contribution to establish that standard for regulated financial services. It governs agents across their full lifecycle — identity, authorisation, behaviour monitoring, and how they interact with each other — within a single architecture.”
To understand why this matters in practice, take something as fundamental as identity. When a human leaves an organisation, their access is revoked. When an AI agent completes a transaction, its identity and permissions do not automatically expire. It can persist in a system long after its mandate has lapsed — with no verified identity anchor, no accountability chain, and no mechanism to intervene. “Also, the longitudinal behavioural trail, if without safeguards such as time limits or privacy protection, risks being tracked and exploited. Hence, holistic lifecycle governance is imperative,” she added.
The Governance Gap Agentic Finance Has Yet to Close
Financial institutions globally are deploying AI agents to initiate payments, execute compliance decisions, and manage portfolios, yet fewer than one in three organisations have adequate governance and controls in place to oversee them, according to McKinsey’s 2026 State of AI Trust survey. Similarly, PwC’s Global AI Performance Study 2026, found that while Singapore businesses outperform the global average on AI adoption (67 per cent report a higher risk appetite for AI investment versus 41 per cent globally), only 47 per cent have a documented responsible AI framework, compared to 63 per cent among global AI leaders.
In January 2026, Singapore’s Infocomm Media Development Authority (IMDA) published the an – the world’s first cross-sector governance framework for AI agents. Budget 2026 built on this with the establishment of a National AI Council chaired by Prime Minister Lawrence Wong, designating finance as one of four national AI mission sectors and committing to regulatory sandboxes for AI innovation.
Ms Tin Pei Ling added:
“We developed KYA drawing on IMDA’s Model AI Governance Framework for Agentic AI, and we went back to IMDA directly to seek their feedback. We are in active engagement with other regulators and stakeholders. We are not presenting this as a finished answer. We are publishing it openly, because this is not a problem any one institution can resolve on its own. We are asking financial institutions, regulators, and technology partners to adopt it, challenge it, and build on it with us.”
To the best of MetaComp’s knowledge and based on publicly available information, no licensed financial institution has published a governance architecture addressing agent identity, authorisation, action scope, behavioural monitoring, risk scoring, audit trails, and agent-to-agent governance in a single framework specifically for regulated payments, compliance, and wealth¹ workflows.
KYA governs AI agents across their full operational lifecycle, to establish who the agent is, what it is permitted to do, what it actually does and how it interacts. The framework is organised across four pillars: Agent Identity and Registration; Authority and Permission Control; VisionX Behaviour Monitoring and Risk Intelligence; and Ecosystem and Interaction Governance, which extends the FATF Travel Rule to agent-to-agent transactions.
Under the KYA framework, every AI agent is anchored to a verified identity linked to a real-world individual or institution through a tamper-resistant registry, ensuring clear accountability from the outset. Each agent operates within strictly defined permissions – governing what it can access, decide, and execute – with built-in safeguards that require human escalation when actions exceed approved thresholds.
The framework goes beyond traditional controls by introducing continuous, real-time monitoring of agent behaviour, assessing not just what actions are taken, but how they are executed and whether outcomes align with intent. As agents operate, their risk profiles are dynamically updated, enabling proactive risk management. All activities and interactions are securely authenticated and recorded, creating a comprehensive, end-to-end audit trail that delivers full transparency and traceability for regulators, institutions, and ecosystem participants.
KYA extends its governance to agent-to-agent interactions, building on the principles of the Financial Action Task Force (FATF) Travel Rule by requiring the exchange of verified identity and transaction information not only between institutions, but also across agent-initiated and agent-to-agent activities within a unified architecture. This ensures that every interaction remains traceable, attributable, and compliant by design.
The framework governs all agents operating within the StableX Network, including those accessing MetaComp’s capabilities through the AgentX Skill ecosystem. Financial institutions and developers can access MetaComp’s regulated infrastructure (compliance, payments, and wealth¹ management) directly through the AI platforms they already use, including Claude, Claude Code, and other compatible platforms via Model Context Protocol (MCP).
The ecosystem’s first Skill, the VisionX Know Your Transaction (KYT) Skill, packages the Web2.5 VisionX Engine into a single agent-callable compliance layer combining more than four blockchain analytics vendors in parallel. New Skills across cross-border payments, treasury, and wealth[1] management will be available by late Q2 2026.
The Compliance Foundation and the Evidence Behind It
The framework sits on top of a compliance architecture that MetaComp has validated across real-world transaction flows. Cross-border transactions today increasingly span both traditional banking rails and blockchain networks within a single transfer. FATF data from June 2025 shows that 73 per cent of jurisdictions have passed Travel Rule legislation, but 59 per cent have taken no supervisory or enforcement action.
Ms Summer Yu, Group Chief Compliance Officer, Alpha Ladder Group, said: “Today’s compliance frameworks were designed for a world where humans initiate transactions. That assumption no longer holds. Our analysis of more than 7,000 real-world transactions shows that even in hybrid fiat and blockchain environments, relying on a single screening tool can leave up to 25% of high-risk exposures undetected. In an agent-driven environment, these risks multiply, and without a defined identity layer, clear authorisation boundaries, or shared accountability standards, the control framework simply does not exist. VisionX Web2.5 closes the visibility gap. KYA establishes the governance layer. Both are essential, and both must be in place before agentic finance can scale safely.”
Today’s announcement continues a period of sustained momentum for MetaComp. Since closing US$35 million across two Pre-A funding rounds within three months, the group has launched the Web2.5 VisionX Engine, established a joint venture with Maqam International Holding to connect Abu Dhabi’s real asset base to Asian capital markets through the StableX Network, and now introduced the KYA Framework as the governance layer for the next phase of institutional agentic finance. Capital is being deployed across all three dimensions: deepening compliance capabilities, expanding regulated payment corridors across Asia, the Middle East, Africa, and Latin America, and building the institutional standards that the agentic era requires.
¹ All products and/or services in relation to securities and capital market products are offered and operated solely by Alpha Ladder Finance Pte. Ltd.
在 Money20/20 亚洲 峰会上发布的 StableX “了解您的代理（KYA）” 框架，明确了为支付、合规、财富管理等金融场景提供服务的 AI 智能体，应如何完成身份识别、授权控制、行为检测与责任归属机制 —— 该框架由新加坡团队制定，面向全球应用
泰国曼谷 2026 年 4 月 21 日 – MetaComp Pte. Ltd.（简称 MetaComp）今日正式发布 “StableX 了解您的代理（KYA）框架”。该框架专为在支付、合规及财富¹管理等受监管金融服务中运行的 AI智能体而设计，系统性建立其身份识别、授权控制、行为监测与责任归属机制，由持牌金融机构主导制定，具备行业开创性意义。MetaComp 是亚洲领先的统一 Web2.5 一体化数字金融解决方案提供商，通过集团级平台打通法定货币与稳定币能力，覆盖支付、资金管理及财富¹管理领域。KYA 框架将向金融机构、监管及网络合作伙伴开放，推动 AI 智能体理在受监管金融体系中的安全应用与规模化发展。
本次发布于泰国曼谷举行的 Money20/20 亚洲峰会同步公布，同时MetaComp还宣布拓展 其 AgentX 代理金融服务 Skill 生态系统。这是持牌金融机构推出的业内首个同类生态，自 2026 年 4 月 21 日起，可在 Claude、Claude Code、OpenClaw 及其他兼容 AI 平台使用，访问官网 www.metacomp.ai 即可体验。
MetaComp 联系总裁 Tin Pei Ling 女士表示：
“AI 智能体已经在金融服务中实际应用——发起支付、做出合规决策、管理投资组合。但目前，业界尚未形成公认标准，来界定这些代理的身份、被允许做什么，以及当其超出授权范围行事时由谁负责。KYA 正是我们在受监管金融服务中对这一空白的积极回应。它在单一架构内，对代理的完整生命周期进行治理——涵盖身份、授权、行为监测以及代理之间的交互机制。”
要理解此举的现实意义，可以从身份这一基础问题说起。当一个员工离开某个组织时，其系统访问权限会被撤销。但当一个 AI 智能体完成一笔交易后，其身份和权限并不会自动失效。代理可能在授权期限结束后仍长期存在于系统中——没有可验证的身份锚点，没有责任追溯链条，也没有干预机制。此外，若缺乏时限、隐私保护等安全保障，AI 智能体的长期行为轨迹还可能被追踪和利用。因此，全生命周期的治理至关重要。”她补充道。
代理金融尚未填补的治理缺口
全球金融机构正在广发部署 AI 智能体，用来发起支付、执行合规决策和管理投资组合。但麦肯锡 2026 年 3 月发布的《AI 信任状况》调查显示，，不足三分之一的机构建立了完备的治理与管控机制对其进行监督普华永道《2026 年全球 AI 效能研究》同样指出，新加坡企业的 AI 应用率高于全球平均水平（67% 的新加坡企业对 AI 投资的风险承受能力更高，而全球平均水平为 41%），但只有 47% 的新加坡企业形成了成文的负责任 AI 框架，而全球 AI 领导者的这一比例为 63%。
2026 年 1 月，新加坡资讯通信媒体发展局（IMDA）发布了《代理式 AI 模型 AI 治理框架》，为全球首个跨行业的 AI 智能体治理框架。2026 年财政预算案在此基础上进一步推进，宣布成立由黄循财总理担任主席的国家人工智能委员会，将金融领域列为四大国家 AI 重点任务领域之一，并承诺为 AI 创新设立监管沙盒。
MetaComp 联系总裁 Tin Pei Ling 女士表示：
“我们借鉴 IMDA 的《代理式 AI 模型 AI 治理框架》开发了 KYA治理框架，并直接向 IMDA 征求意见。目前我们也在与其他监管机构和各利益相关方积极沟通。我们并非将此作为一个终极答案，而是选择公开发布，因为这一问题并非单一机构能够独自解决。我们呼吁金融机构、监管机构和技术合作伙伴共同采用、提出改进建议，并与我们一起完善这套体系。”
据 MetaComp 所知及公开可获得的信息，目前尚无持牌金融机构发布过在一个单一框架内同时涵盖代理身份、授权、行动范围、行为监控、风险评分、审计追踪以及代理间交互治理的治理架构，并专门针对受监管的支付、合规和财富¹等金融服务场景。
KYA 框架对 AI 智能体实现全运营生命周期治理，明确界定代理是谁、被允许做什么、实际做了什么以及如何交互。该框架围绕四大支柱构建：“代理身份与注册”；“授权与权限控制”；“基于VisionX的 行为监控与风险智能”；以及“生态系统与交互治理”。其中，交互治理进一步将 FATF 旅行规则的理念延伸至代理与代理之间的交互场景，实现可追溯、可验证的合规机制。
在 KYA 框架下，每个 AI 智能体都将绑定经核验的身份，并通过防篡改登记系统关联至现实世界中的自然人或机构，从源头确保清晰可追溯的责任归属。所有智能体均在严格限定的权限内运行，明确其可访问、可决策、可执行的范围；框架内置安全机制，一旦操作超出预设阈值，将自动触发人工介入。
该框架在传统控制措施之上进一步延展，引入对 AI 智能体行为的持续、实时监测机制。其评估维度不仅覆盖代理“做了什么”，还包括“如何执行”以及“结果是否符合预期意图”，从而实现对行为全过程的立体化审视。随着代理的持续运行，其风险画像将被动态更新，使风险管理由被动响应转向主动识别与干预。同时，所有活动与交互均经过安全认证并完整记录，形成覆盖全流程的端到端审计追踪，为监管机构、金融机构及生态参与方提供高度透明且可追溯的合规基础设施。
KYA 框架将治理范围延伸至AI 智能体之间交互场景，并基于金融行动特别工作组（FATF）旅行规则的核心原则作出强化：在统一框架下，不仅要求机构间交换经核验的身份与交易信息，同时覆盖代理发起、代理交互的全流程。这一设计从底层确保每一次交互均可追溯、可归责、合规可控。
该框架适用于 StableX Network 内运行的所有代理，包括通过 AgentX Skill 生态系统调用 MetaComp 服务能力的代理。金融机构于开发者可通过日常使用的 AI 平台（包括 Claude、Claude Code 以及其他通过模型上下文协议兼容的平台）直接接入 MetaComp 的受监管基础设施（合规、支付和财富¹管理）。
该生态系统的首个 Skill——VisionX 交易核查（KYT）Skill——将 Web2.5 架构的VisionX Engine 封装为可由AI 智能体直接调用的合规层，可并行对接超过四家区块链分析供应商的数据。面向跨境支付、资金管理和财富¹管理的更多新 Skills 将于 2026 年第二季度末前推出。
合规基础与实践依据
该框架建立在 MetaComp 已通过真实交易流程验证的合规架构之上。当前跨境交易往往在一笔转账中同时涉及传统银行通道和区块链网络。FATF 2025 年 6 月的数据显示，全球73% 的司法管辖区已出台旅行规则相关立法，但 59% 尚未采取任何监督或执法行动。
Alpha Ladder 集团首席合规官 Summer Yu 女士表示：
“现行的合规框架，是建立在由人类发起交易这一前提之上的。然而，这一假设已不再成立。我们对超过 7,000 笔真实交易的分析显示，即使在法定货币与区块链交织的环境中，依赖单一筛查工具，可能导致高达 25% 的高风险暴露被遗漏。而在由 AI 智能体驱动的场景中，这类风险将被进一步放大。如果缺乏清晰的身份层、明确的授权边界以及统一的问责标准，所谓的控制框架将难以成立。VisionX Web2.5 解决的是可见性问题，弥补跨 Web2 与 Web3 的风险盲区；而 KYA 构建的是治理层，确立代理在金融体系中的身份、权限与责任机制。二者缺一不可，且必须同步建立，才能让“代理金融”在受监管体系中实现安全、可持续的规模化发展。”
本次发布延续了 MetaComp 持续增长的发展势头。三个月内，集团先后完成两轮 Pre-A 轮融资，累计融资 3,500 万美元；此后推出 Web2.5 VisionX Engine，与 Maqam International Holding 成立合资企业，通过 StableX Network 将阿布扎比的实物资产与亚洲资本市场连通；如今又发布 KYA 框架，为机构AI 智能体金融下一阶段搭建治理层。资金将投向三大方：深化合规能力、扩展覆盖亚洲、中东、非洲和拉美地区的受监管支付通道，以及构建AI 智能体时代所需的机构级标准。
¹ 所有与证券及资本市场产品相关的产品及/或服务均由 Alpha Ladder Finance Pte. Ltd. 单独提供和运营。
关于 MetaComp
MetaComp 是亚洲领先的统一 Web2.5 数字金融解决方案提供商，通过机构级集团平台整合法币与稳定币能力，涵盖支付、资金管理及财富管理等核心金融服务。MetaComp 持有新加坡金融管理局（Monetary Authority of Singapore，MAS）颁发的主要支付机构牌照（Major Payment Institution），可提供数字支付代币（Digital Payment Token，DPT）及跨境汇款（Cross-border Money Transfer，CBMT）服务，目前服务超过 1,000 家机构及合格投资者客户，业务网络覆盖全球主要金融中心。
2025 年，该集团级平台处理的支付及场外交易（OTC）总量已超过 100 亿美元，覆盖 13+ 种稳定币，月度运行规模超过 10 亿美元。通过 StableX Network（StableX 网络），机构客户可在统一且合规的 Web2.5 金融架构下，实现法币与稳定币资金的转移、兑换及资金管理。
资金与投资服务由 MetaComp 关联公司 Alpha Ladder Finance Pte. Ltd. 提供。该公司持有新加坡金融管理局颁发的资本市场服务牌照（Capital Markets Services，CMS）及认可市场运营商牌照（Recognised Market Operator，RMO），其财富解决方案的管理资产规模已突破 5 亿美元。
截至目前，MetaComp 已完成 3,500 万美元的 Pre-A 轮融资，并于 2025 年实现全年净利润，体现了机构市场对受监管 Web2.5 金融解决方案的强劲需求。
了解更多信息，请访问 www.mce.sg，或在 LinkedIn（https://www.linkedin.com/company/metacompsg） 及 X 平台关注 MetaComp：@MetaCompHQ。