A multitude of players in the banking and capital markets, including Swift, DTCC, and UBS, have successfully completed the second phase of a project led by Chainlink, aimed at leveraging AI and blockchain technology to standardize and enhance corporate actions processing.
Corporate actions processing incurs an annual cost of approximately $58 billion for the global financial sector, with expenses rising by 10% year-over-year and automation rates remaining under 40%, as noted by Chainlink. The 2025 Asset Servicing report from Citi reveals that an average corporate action event involves over 110,000 firm interactions and costs around $34 million to process. Surprisingly, 75% of market participants still rely on manual data revalidation.
In a collaborative effort last year, Chainlink, Swift, Euroclear, and six financial institutions showcased the capability of large language models—such as OpenAI’s GPT series and Google’s Gemini series—to extract structured data from unstructured corporate action announcements, creating unified golden records on-chain. This effectively provides a single source of truth accessible, verifiable, and usable by all stakeholders.
During the second phase of the initiative, 24 organizations—including ANZ, DBS Bank, and TMX—implemented a production-grade deployment. This enhancement significantly improved the speed, reach, and accessibility of corporate actions data, with potential savings of tens of billions of dollars for the global financial system.
Testing involved the Chainlink Runtime Environment orchestrating the validation of various AI model outputs, which were then transformed into ISO 20022-compliant messages for transmission via the Swift Network. Simultaneously, the Chainlink Cross-Chain Interoperability Protocol distributed these confirmed records across DTCC’s blockchain ecosystem and other public and private blockchain environments, enabling simultaneous access across traditional and blockchain infrastructures.
The initiative also introduced designated roles for data attestors and contributors, allowing them to cryptographically verify data accuracy and fill in any missing data, thus maintaining a verifiable chain of custody throughout the lifecycle of each corporate action.
Throughout the testing phase, the system achieved nearly 100% data consensus among AI models for all evaluated corporate actions. Its architecture demonstrated support for multilingual processing, accommodating disclosures in non-English languages such as Spanish and Chinese.
Chainlink asserts that this system offers a unified golden record for corporate actions—a verifiable, real-time source of truth accessible to smart contracts, custodians, and post-trade systems. This infrastructure also facilitates tokenized equities, an increasingly popular category of tokenized assets, allowing them to reference the same confirmed records across both public and private blockchains. This advancement lays the groundwork for improved synchronization and greater automation within on-chain markets.
Looking ahead, the group plans to extend the workflow to encompass more complex corporate actions, such as stock splits, broaden global coverage with support for additional jurisdictions and currencies, and implement enhanced privacy and governance controls.
Nigel Dobson, Lead of Banking Services at ANZ, commented: “The Chainlink-led initiative exemplifies how blockchain can provide attested, real-time corporate actions data across various jurisdictions and platforms, reducing operational risk, enabling smart contract automation, and establishing a foundation for scalable digital asset servicing.”