Trends towards AI-supported Modular Knowledge Management

In this second article in [A]’s series on Knowledge Management, we focus on the importance of knowledge itself, and approaches to improving the free-flow of knowledge throughout an enterprise.

In our white paper, “Creating Content-as-a-Service through smart Knowledge Management practices,” we discuss why Knowledge Management’s biggest value is still to be fully realized. We see how the combined impact of engineering content sets, graph technology, and Content-as-a-Service architectures are setting the course to innovate enterprise Knowledge Management. Unhindered access to effective knowledge will remain impaired until content orchestration models and methodology move beyond the traditional siloed workflows.

Forces that are evolving content workflows

Today’s world of rapidly multiplying content assets features ever-changing channels for publishing. Unprecedented ways of thinking and working are emerging to address these new markets and opportunities. Knowledge cannot efficiently serve consumers in only traditional document form. Creating and managing knowledge assets the ‘old way but faster’ is not sustainable. As a result, the shape of knowledge – information – the content itself – is evolving towards modular forms that can react to a customer’s unique circumstances, contexts, and experiences.

Since the shift to this new world has already happened, the migration away from single-silo Knowledge Management applications is a foregone conclusion. Today organizations are rapidly navigating towards an AI-supported and extended organizational “knowledge sphere”. Converting knowledge capital into structured, semantically-rich forms, and connecting those assets to machine learning and other AI enhancements are essential for survival in this new world.

[A] observes the following three trends that should inform any progress on Knowledge Management strategy in the near future:
 

• Intelligent Content: Enterprises are now migrating to component-based content for both Knowledge Management systems and customer experience (CX), in order to fuel the assembled, contextual use of content. 

 

• Semantic Services: Organizational approaches are now heading towards modular and graph-based semantics that can be made available via service-provider relationships to multiple systems and applications across teams and platforms.

 

• Content-as-a-Service (CaaS): Enterprises are moving away from a siloed publishing workflow and towards the ability to deliver modular content assets via APIs to many downstream users and applications.

More about these trends

Some tenets of content intelligence include:
 

• To reach humans and robots, published content must be intelligent.
• Intelligent content becomes more valuable as humans and robots interact with it.
• The more interaction, the more likely that content will influence as a result of human and robot impressions of the content. Interactions equal impressions.
• Impressions are enabled by structure and driven by content reuse.

A semantic services platform enables new, key organizational capabilities, including:
 

• Content targeting and personalization
• Faceted, intelligent search experiences
• A rich corpus of content to aid chatbots and other conversational UIs
• Discovery of content assets for reuse and multipurpose value
• Reduce internal friction in creating consistent content across divisions

Characteristics of content-as-a-service solutions include:
 

• Content provided out to the world via a REST-based API
• An approach to architecting content within defined content models
• Structured formats for returning content via simple queries
• Distributed authoring and workflow content administration
• A content repository hosted in the cloud for universal access
• Triggers that alert customer experience applications that consume content to content updates
• Metadata definitions that can be defined and move along with the content via API

As a result of these trends, content intelligence continues to grow across enterprises, as new approaches to content get introduced, demonstrate value, and continue to expand adoption.

Artificial Intelligence: an incomplete solution without Content Intelligence 

Given the vast stores of information surrounding modern organizations, it’s natural to seek knowledge management through automation. Vendors promise that content can be automatically indexed and combined across disparate platforms through natural language processing (NLP), entity recognition, and even created into new forms with natural language generation (NLG).

Nevertheless, any AI service that attempts to process unstructured content runs into significant error rates. Machines have a hard time disambiguating complex human communications. As any IBM Watson customer can attest, training machine models on unstructured content is extremely human-capital intensive and leads to highly imperfect results. Adopting Knowledge Management practices and a proper Content Services Organization (CSO) can greatly improve this process and results.

As Alan Porter wrote in AI's Missing Ingredient: Intelligent Content: 

“We are carrying a lot of technology debt that is hampering the truly seamless AI customer experience. When the systems we already have don’t interact, and companies continue to build point solution silos, duplicate processes across business units, or fail to take a holistic view of their data, content, and technology assets, then the existing AI systems will continue to pull from a restricted set of information.”

Benefits of Structure and Semantics

The use of semantics and structure transforms these systems from complicated, siloed, and difficult to use, to user-friendly and efficient solutions. They are both essential for content that can be understood by machines and humans.
  Content structure shapes content with an object-oriented approach:
 

• Content is organized as reusable content objects, rather than hard-coded, unstructured blobs
• Objects have containers, which can be manipulated, transformed, annotated, reused, and managed from a central location; they can be pointed to from anywhere

Content semantics is the contextualization of content structures:
 

• Content semantics define the entities, associations, and relationships for a given piece of content within the metadata
• Semantics is how machines can understand content connections and relationships 

Eventually, AI will come to play a pivotal role in constructing and comprehending the insights within rich, ontologically-sound knowledge graphs that power everything. Knowledge management will become amplified in regards to effectiveness by machine-assisted learning, accelerating enterprise knowledge value, and availability to human and machine consumers.