In the early days of the internet, efforts to digitize physical information focused on creating scanned copies and PDFs of paper records. This approach, spanning industries from newspapers to eBooks, education, finance, and online shopping, soon proved inadequate. Readers found the content difficult to navigate and read, prompting businesses to adopt more digital-friendly formats.
Newspapers began producing searchable articles, dynamic content, and interactive features. In education, digital learning platforms emerged and there was a shift towards video lectures, interactive quizzes, and real-time feedback. E-books evolved to include features like adjustable text sizes, hyperlinks, and multimedia content, leading to the development of XML-based formats like EPUB.
The data garnered allowed businesses to enhance user experiences and outcomes.
For rule-makers, paper has several benefits that have been difficult to replicate digitally such as fixity – a property that ensures that a document can be relied upon as the master version. In addition, there are historic and traditional processes rooted in paper that may not easily be replicated in digital environments and require a specialized approach.
The creation and consumption of rules are increasingly digital, meaning that computers are now prevalent. Businesses across the board are moving past the PDF and early approaches to digitizing content – the digital proxies for paper. Transitioning to XML schemas and formats is often thought of as the solution. However, XML in and of itself is not digital transformation. XML is one piece of a larger puzzle.
Knowledge graphs are emerging as a critical tool for integrating data-first approaches.
The value of data-first approaches to rulemaking
There exists a mental model where rules are perceived to be ‘read’ linearly. However, in reality, today’s businesses are chiefly concerned with managing the domino effect of change that updates to the rules trigger – from the knowledge systems to the learning management systems used to train staff and the ‘cheat sheets’ on the floor. These businesses require greater ability to plug the requirements – the applicable changes – directly into their world. This typically involves copying and pasting relevant text.
Moving to a data-first approach has the potential to deliver outcomes for rule-makers including:
- Tighter relationships with the end-consumer
- Transparency of process
- Ability to create value and strengthen the position of the rules in society into the future
- Increased ability to deliver added value and richer experiences
Transitioning from document- to data-first approaches doesn't mean sacrificing the process.
Knowledge graphs: reducing the risk of AI hallucinations
Knowledge graphs are emerging as a critical tool for integrating data-first approaches, especially as generative artificial intelligence (AI) reshapes industries. Knowledge graphs ground Large Language Models (LLMs) with their ability to represent both structured and unstructured data – a technique known as retrieval augmented generation (RAG).
Knowledge graphs provide a structured way for LLMs to understand the relationship between pieces of content and improve contextual understanding, thus reducing the risk of hallucinations. Knowledge graphs are also an opportunity for rule-makers to produce AI assets.
To achieve RAG LLMs, however, rule-makers must first address their approach to content.
Structured content: beyond pagination-based approaches to content
Component-based, structured content models are an ideal fit for rules. A component may represent a requirement, a clause, a measurement parameter, a procedure, etc., which reflects how consumers think about standards.
Structured content allows content to be managed at a micro level, enriching fragments with metadata and tagging to enhance its downstream use. With the help of technology, structured content allows rule-makers to produce knowledge graphs.
Preserving the process
Transitioning from document- to data-first approaches doesn’t mean sacrificing the process. Indeed, addressing the digital transformation of rules is critical to enhancing and safeguarding rule-making processes into the future.
Both rule-makers and rule-takers require increased visibility and traceability over standards. Rule-makers require increased means of delivering value and creating tighter relationships with their customers. This will only become more apparent as technology advances.
Moving to a data-first approach to content necessitates an overhaul in technology, however, crucially, this should not result in tooling that alienates professionals and subject matter experts. Rather, technology should meet experts where they are, leveraging technology to enhance the process rather than demanding process change for the sake of technology.
