Post your ideas for graph-powered content applicaitons here

[FirstLight]

The sky is the limit when it comes to the potential use of knowledge graphs with our TechComm and MarComm content. Let's gather them here in this thread.

Regards,
Michael

[FirstLight]

wally.jpg (45.97 KiB) Viewed 10497 times

I don't get it Wally, what do you mean by graph-powered content applications?
Well Beaver, the folks that invented structured content - what they now call Intelligent Content nowadays - didn't expect that the only value organizations would get was reusable and repurpose-able content, single sourcing, consistency, device- and presentation-independent output, omnichannel publishing, you know - all that write-once-publish-many stuff.

No Wally, what else did they expect?

Well, Beav, they began adding extra junk in there called sematic intelligence. They knew that machines would one day be able to use it to create amazing applications that weren’t even possible back then but can today. It would take about 40 more years, now with artificial intelligence and machine learning came of age.


That sounds neat Wally! So, what kind of applications are possible?


A whole slew of ‘em Beav. Here’s just a few:

• For starters, content search will be exponentially more accurate and relevant over traditional indexed search - we'll get that with taxonomy-enriched content, but it will get even better with knowledge graphs
• Faceted search will be much easier to implement and use
• It will help us create and maintain content models
• Knowledge graphs will provide new insights into all of the content. We can query our graphs to:
  • Discover gaps in our content
  • Identify conflicts and duplication
  • Understand how all of our content fits interacts - product, support, marketing, partner, developer, sales, learning

For example, an elusive goal has been identifying how our post-sales content contributes to revenue. We can query content usage and sales if we extend our graph to sales intelligence/ Finally, we can prove to C-Suite types what many of us have known for a long time - that post-sales content contributes to conversions every bit as pre-sales content - often at a higher rate. Yes investment in post-sales content is significantly less than Marketing because it is mistakenly viewed as a cost center.

Wow Wally! What else?

That's just for starters Beav!

• Our content management system will be able to recommend the right content to the user at the right time.
• In fact, the AI/ML will be able to go get lots of little chunks of content from all over the place automatically organize and present it
• When some need support, the customer will be able to get solutions to their issues without a human support agent.
• When a human support agent is needed, that agent will have fast and accurate access to the content he or she needs to help solve the customer's issue.
• Even better, when support agents solve issues and document them, that content will find its way back into the main product content so that other customers have an easier time of it. In fact, machine learning will eventually get so good, the content will start to fix itself. That's what's called autonomic content.

Wooahha Wally! That could like the content almost going to write itself?

That's already happening Beav. Look up jasper and try it out. It will be a very long time before it gets really good, and it won't replace writers.\ They'll become content designers. The kind of semantic intelligence I'm talking about will make that much better.

...and that's just the tip of the iceberg.

[ClaudetteH]

Hey folks! I'm super excited about our discussions, but have some baseline questions that will help me get a sense of our scope during discussions:

Is everyone's content publicly accessible (i.e., no login required)? I ask because it seems like content that's publicly facing will need to adhere to a variety of standards (DITA, SEO-optimized, etc.) in addition to any cool thing we can cook up.

Are folks primarily concerned about tech docs? Or does their scope include things like training content, marketing docs, knowledge base articles, community forums, etc.?

Are we primarily thinking about written content? Or are we also thinking of a classification system for images, videos, podcasts and other multimedia?

Thanks!
Claudette

[FirstLight]

The Avalara content lake will contain publishable content from all content domains (product doc, enablement (support, training, sales enablement, developer, partner, and more), and all types of content (text, videos, PPts, PDFs, DITA, pretty much anything that is a deliverable object. Again, it will be the taxonomy classification for each object that governs who can see what.

An org doesn't need to federate content from multiple domains to do this - it's a big long-term lift fr many to deliver on busting of content silos, but it can start with just one domain (product content for example) and then be additive on e domain at a time (and usually advisable.

The real key to success is the establish a common enterprise content taxonomy and map the content types models to align them.

Interested in others' responses too.

Michael

[ClaudetteH]

Love that answer, Michael. You are describing the dream for me. Break down the silos!

[JohnO]

Hello cont-graphers

What if there was a way to identify, classify, and connect business lexicons, meta-, master-, and reference-data to content in a graph via metadata? Would hearing more about this idea appeal to graphistas on this thread?

The approach is technology agnostic, language neutral, and domain independent and employs a cloud-based CMS to 'get to the content'.

I spoke with Michael this morning and I think by his level of interest that there just might be!

John O'
CEO and Co-Founder of Semantium

[FirstLight]

John,

I never replied to your post - YES!!! An emphatic YES! To do this we have to align (at least) our internal enterprise terminology, shared taxonomies, and domain ontologies. As we do that and create various (specific domain) graphs across functional areas (business ops, finance, planning, engineering/architecture as well as our content - we can then join the various graphs. In essence, we then have an enterprise knowledge graph (EKG) that can be queried via graph query languages such as SPARQL to construct new and novel applications no one can possibly do now with current RDB technology. Imagine combining your EKG with tech in the class of ChatGPT! ChatGPT is good, but combined with domain-specific intelligence it would be like putting it on super steroids. More folks need to understand the technology behind ChatGPT and what it can and can't do on its own. We could take it even further with external assets via something like iiRDS (Feb SCGG topic) I think maybe it is time to plan a SCGG session on ChatGPT this Spring, 'eh?

Regards,
Michael