From PDFs to JSON: How Chainlink Is Solving One of the Biggest Problems in Finance

At #Sibos 2024, Swift’s Kelli West described corporate actions as “one of the most complex and costly unstructured data problems in the financial world.”

Our recent industry initiative is set to change that.

Today’s corporate action events—dividends, stock splits, mergers, and more—are typically shared through PDFs and press releases, ultimately fragmented, unstructured data. Managing this data is a messy process that costs institutions millions.

It doesn’t have to. Large language models (LLMs) can easily extract critical data from these PDFs and press releases, reducing the messy manual work required. However, LLMs can sometimes hallucinate.

This is where Chainlink’s decentralized oracle networks (DONs) come in. DONs request data from multiple LLMs and then come to a consensus on it, addressing hallucination risks. Next, they deliver the data in a machine-readable, structured format, i.e., JSON.

But this is just the start. Unstructured data is just one data problem in finance today.

Another is the lack of a single source of truth, which results in firms repeating the same data-cleaning efforts at each step as data is passed from public companies to brokers and other intermediaries and ultimately to investors.

To overcome the lack of a single source of truth, structured data can be stored on blockchains as a unified golden record that everyone from asset issuers to individual investors can easily access via Chainlink.

Not only can this help institutions reduce costs by minimizing reconciliation processes, but it opens up deeper efficiencies and innovation by allowing programmable actions based on corporate actions data to be automated.

Get the full story: