Mike's Notes
Good words from John Gorman in Canada.
Resources
References
- Reference
Repository
- Home > Ajabbi Research > Library >
- Home > Handbook >
Last Updated
04/12/2025
Knowledge Graphs Should Not Be Just for Analytics and Insights
Information management professional specializing in semantic interoperability, with over 30 years of experience. Principal, Founder and Chief Disambiguation Officer at Quantum Semantics Inc in Calgary. Inventor of the Q6 information management model.
There is, in my view, a large and currently under-served audience for knowledge graphs in the enterprise. These are employees, contractors, and yes, even executives who need access to a more granular level of access to company knowledge.
Maybe we should start calling them 'Learning Graphs' instead.
The interesting thing is, by providing this level of service, solutions to other gnarly challenges like data governance, metadata, reference and master data management, data fluency, semantic interoperability, and when done properly (i.e. FAIR-From-Birth) dimensional analysis of operational data stores, make themselves available.
Sounds ambitious? Why not? Enterprise Information is now all about language, so how about we start there and leverage persistent patterns of classification and usage. Here are some of the immediate and 'downstream' benefits:
- Data Governance. Most companies make the mistake of starting with data. The more natural approach - and with a lot less risk to life and limb - is to begin with the language of the business. And, you get to roll out a 'FAIR-From-Birth' business glossary as a side benefit, just like the big boys and girls.
- Metadata. Business users rarely even care about metadata, but when they see how the language of the business connects to and is equivalent to metadata values the light bulbs start to go on. As a bonus, they also get to see how missing, misspelled, and misshapen vocabulary gums things up.
- Reference and Master Data. This is another opportunity to see how the information supply chain should work. Crap components upstream means crap assemblies downstream. Ask your local supermarket manager how he handles a shipment of rotting tomatoes.
The benefit for business owners is two-fold:
- They get to see what kind of cleanup is required when they throw crap over the fence.
- They also get to see how one-and-done actually reduces their workload.
Data Fluency. Seeing the connections between how they talk about the business and what kinds of data those conversations connect to? Priceless.
Semantic Interoperability. Ah, yes... the Holy Grail. What if we made it possible to access semantically equivalent pairs as a start? So, if Jane McCallum, the CFO of Acme Inc. doesn't know (have knowledge of) the meaning of the acronym FLOC, she can simply look it up on her phone during her Monday executive team meeting and learn it. No embarrassing interruptions, just immediate access to enterprise knowledge. What a concept!
Analytics and Insight. Finally, the raison d'être for almost every technical innovation so far this decade. When an information supply chain starts with the notion that every component, especially the very granular ones, should be engineered to fit into a downstream, multi-dimensional ecosystem of semantically connected information good things start to happen.
DM me if you want to learn more about Semantium's set of protocols.
No comments:
Post a Comment