Building Out the Enterprise Knowledge Graph

There is this notion of the "enterprise knowledge graph".  That term "enterprise knowledge graph" and the base term "knowledge graph" is very overloaded these days.  If you asked 10 people, you would get 10 different answers.

Here is Google's definition of "enterprise knowledge graph": "Enterprise Knowledge Graph organizes siloed information into organizational knowledge, which involves consolidating, standardizing, and reconciling data in an efficient and useful way."

Per Object Management Group (OMG) an enterprise knowledge graph is: "Enterprise Knowledge Graph (EKG) represents the integration of information and knowledge of an enterprise and its ecosystem."

Per the Enterprise Knowledge Graph Foundation (EKGF) an enterprise knowledge graph is: "An EKG is a semantics-first foundation for an enterprise: it defines context (concepts, relationships, rules) and connects that meaning to high-quality, reusable facts and data products—so systems can interpret data consistently. In practice, it acts as a governed semantic layer in front of your internal and relevant external systems."

Per the consultancy Enterprise Knowledge an enterprise knowledge graph is: "An enterprise knowledge graph is a representation of an organization’s knowledge domain and artifacts that is understood by both humans and machines. It is a collection of references to your organization’s knowledge assets, content, and data that leverages a data model to describe the people, places, and things and how they are related."

Per the technology company Quinnox; Gartner describes an enterprise knowledge graph thus: "Knowledge graphs are 'graph-based data structures that capture the semantics and relationships among data to support enhanced context, insight, and data-driven decision-making.' By 2026, Gartner predicts that enterprises using semantic and graph-based approaches will reduce artificial intelligence technical debt by 75% compared to those relying on traditional architectures."

The Business Rules Community (BRC) has no formal definition of an “Enterprise Knowledge Graph”, but it does define an equivalent concept: the Business Knowledge Blueprint: "A structured, semantically precise, business‑friendly representation of an organization’s concepts, facts, and rules."

The obvious thing to do is to first break down each word to understand what those individual words mean. Here are my definitions of: enterprise, knowledge, and graph:

  • "Enterprise" is really any economic entity; small, medium, or large. An enterprise is organized and has some purpose or "mission" or "mandate". It could be a business, a not-for-profit organization, a government agency. An enterprise mobilizes people, processes, and resources to achieve some specific goal or produce some sort of value for it's stakeholders at scale. Microsoft is an enterprise; so is the state of Washington, so is the Italian restaurant where I had dinner last night.
  • "Knowledge" is  all accumulated trusted information which has been interpreted, internalized, justified, and understood within the scope of some domain of understanding (a.k.a. community, shared definition; shared understanding) giving the information meaning. In this context, knowledge is what an enterprise knows to be true.
  • A "graph" is a mathematical structure of "nodes" and "edges", explained by graph theory, that models things and the connections between those things. Basically, a graph is an approach, a powerful tool, to storing meaning. That meaning can be stored with unprecedented clarity 

An enterprise knowledge graph is best understood in the context of "artificial intelligence". And I don't mean understanding the definition of artificial intelligence; what I mean is that if an enterprise can effectively harness the power of artificial intelligence, what exactly does that mean for the enterprise?

Why would an enterprise put knowledge in the form of a graph? Why would they do that? Why should an enterprise go through all that effort and what would that graph of enterprise knowledge look like after that task was complete?

To understand the answer to that question, you first have to understand computers.  As I have pointed out, computers are dumb beasts. Computers do four things extremely well:

  • store information reliably and efficiently
  • retrieve information reliably and efficiently
  • process stored information reliably and efficiently, mechanically repeating the same process over and over
  • instantly accessible information, made available to individuals and more importantly other machine-based processes anytime and anywhere on the planet in real time and with today's internet do this cheaply
So, not everything is computable.  But there are four major significant obstacles that must be overcome in order to harness the tremendous power of computers:

  • business professional idiosyncrasies: different business professionals use different terminologies to refer to exactly the same thing
  • information technology idiosyncrasies; information technology professionals use different technology implementation options, techniques, and formats to encode and store, retrieve, and process exactly the same information
  • inconsistent domain understanding: business professionals understand knowledge differently and technology has limitations related to expressing interconnections which many business professionals don't tend to understand well
  • computers are dumb beasts: computers don't understand themselves, the programs they run, or the information that they store, retrieve, process, or provide access to; computers only interpret

Having defined the terms I am using and having laid out some fundamentals that are necessary to comprehend the information I am about to convey; this is my personal take on  the enterprise knowledge graph.

An Enterprise Knowledge Graph is the semantic backbone of a modern enterprise. That enterprise knowledge graph serves as a "blueprint" or "model" of the enterprise which is interpretable by both humans and machines such as computer based processes. The more enterprise knowledge you can put into that enterprise knowledge graph; then the more explicit, implicit, and tacit, knowledge both humans and machines can make use of and the more humans and machines can team up and work together.

This enterprise knowledge graph is a shared understanding.  It takes a tremendous amount of work to create that shared understanding, that shared conceptualization. And that conceptualization is not static, it is dynamic and it must be governed. Epistemic risk, the risk that your understanding is wrong, must be managed effectively.

The enterprise knowledge graph is the scaffolding of the modern enterprise.  If done well, industrial processes can be created to replace what amounts to "bucket brigades" used to run the enterprise.

There are four approaches to creating an enterprise knowledge graph:

  1. One standard enterprise knowledge graph format is created and every enterprise will use that one standard enterprise knowledge graph. This will not tend to work because enterprises have differences.
  2. Each enterprise separately create their own unique enterprise knowledge graph, crafted to their individual needs.  This will not tend to work because (a) it is not very efficient and (b) each individually unique knowledge graph will need unique software to work against that knowledge graph (i.e. does not scale).
  3. One global open industry standard framework is used as a template which can be modified to meet the unique individual needs of industry groups or uniqueness's of enterprises.
  4. A hand full of separate global open industry standard frameworks might be created because sometimes it is hard to agree on one thing and multiple alternatives might prove beneficial.

One standard "form" that every enterprise must then fit into make zero sense because that cannot work.  Each individual enterprise doing their own thing makes little sense, no scale.  Everyone agreeing on one approach might be optimal, but probably not achievable. My personal bet would be on a hand full of frameworks.

Is it really going to be an enterprise knowledge graph? Or, would it be better called an enterprise hypergraph?  There is a difference between a knowledge graph and a hypergraph.  In my personal view, the enterprise knowledge graph will work similar to, say, Google Maps, and have different "layers" which can be viewed.  For example, if you want to see the "accounting layer" you can filter the hypergraph and see that specific aspect.  Whatever it ends up being,  it MUST be readable  by business professionals with a liberal arts degree (i.e. does not require a PhD in computer science).

There are already a lot of standards that exist.  However, these standards were developed in silos.  These silos need to be combined into one unit.  This is an initial  Venn diagram which shows some of the pieces of the enterprise knowledge graph:

Explaining this first cut of what I see within the enterprise knowledge graph, obviously "accounting" and "economic activity" are dimensions "layers" of the enterprise knowledge graph. Enterprises have to publish financial statements.  For example, listed or public enterprises need to provide compliance or regulatory report to their regulators.

I am still trying to figure out precisely what a "semantic layer" is as compared to an "enterprise knowledge graph".  They are not the same thing.  Seems like a sematic layer is a translation layer.  
An enterprise knowledge graph is a meaning layer. They seem to solve different problems, operate at different layers, and have different expressive power. More on that later.

Comments

Post a Comment

Popular posts from this blog

Professional Study Group for Digital Financial Reporting

Image
Digital financial reporting is a thing.  The IFRS Foundation explains digital financial reporting  in an article, Digital Financial Reporting - Facilitating digital comparability and analysis of financial reports . OFFICIAL VERSION : The IFRS Foundation publishes an XBRL taxonomy for IFRS for SMEs (2024).  You can find official information about that XBRL taxonomy here on this web page, IFRS for SMEs as Published by IFRS Foundation (2024) . (Also see this PDF, IFRS Taxonomy Illustrated .) Since that official XBRL taxonomy is effectively a "pick list" created from the perspective of standards setters (i.e. not very practical, easy to use, and intended for ease of use by reporting economic entities);  ENHANCED VERSION : I have created an enhanced version of that same XBRL taxonomy, see  IFRS for SMEs 2024) - ENHANCED (Work in Progress) . The enhancements make the IFRS for SMEs XBRL taxonomy more useful in many different ways.  The enhancements are driven...
Read more

Big Idea for 2025: Semantic Accounting and Audit Working Papers

Image
I began my career in accounting in 1982 as an auditor with Price Waterhouse when accounting and audit working papers and schedules were 100% paper.  But within three months, I was creating many of those same working papers and schedules first in VisiCalc and then Lotus 1-2-3.  I would create those schedules and working papers electronically, then print them out, and tape them into the bundle of paper that supported our audits. The Compaq luggable computer was invented and electronic spreadsheets where even more compelling. Today, accounting and audit working papers tend to be 100% electronic proxies for those paper documents created in Excel or some other electronic spreadsheet, Word or some other word processing document, PDF generated from some word processing document, and maybe HTML.  All of these are presentation oriented proxies for documents and none of them is truly understandable by a computer based process. The next step in the evolution of accounting and audi...
Read more

Rethinking Financial Reporting: the Model-driven Financial Statement

Image
It is time to rethink the design of financial statements. Clay tablets worked well and they had a good run which ended many years ago when the clay tablets were replaced by papyrus and then ultimately paper.  Paper had a good run also but paper financial statements were ultimately replaced by what amounted to an electronic version of that paper; PDF, HTML, and such. We can do better.  Today's demands call for more modern accounting tools . Idealized design  (a.k.a. greenfield project ) is the notion of designing a system without imposing constraints of prior work.  This technique may, or may not, work for rethinking the financial statement. Idealized design might be a harder path in the short term, but taking this approach has very positive long term ramifications. Here are a number of ideas that I would throw at the wall if I were in charge of rethinking the financial statement to make such tools modern: Digital first : The first thing that I would do is abandon the...
Read more