Knowledge management suffered reputation damage across decades despite early promise. Introduced during the 1990s, the discipline experienced popularity cycles interspersed with periods of significant devaluation. However, this characterizes primarily digital incarnations of knowledge management practices.
Knowledge transmission spans centuries through written materials, apprenticeships, formal instruction, training programs, cultural experiences, and folk traditions. Digital knowledge management emerged alongside early collaboration technologies: listservers, online discussions, communities, bulletin boards, and corporate groupware platforms like Lotus Notes and SharePoint.
Connecting people with knowledge and expertise was conceived even earlier. The 1968 demonstration of personal computing concepts, later elaborated through research labs and commercialized by technology companies, planted seeds for contemporary knowledge networks. Now knowledge management experiences revival as its critical role enabling artificial intelligence gains recognition.
Information Access Imperatives
Early business customers articulated fundamental challenges: organizations needed getting more information online. This assessment proved prescient. While universal knowledge access remained distant, content and information explosions across corporate intranets and public internet created overwhelming volumes demanding comprehensive access mechanisms.
Technology companies developed indexing servers enabling information retrieval and expertise location. These early tools evolved into sophisticated platforms, with some underlying DNA ultimately powering advanced cognitive systems. The progression from simple content indexing to complex AI platforms illustrates knowledge management's continuous evolution.
AI's Knowledge Foundation Dependency
Cognitive technology suites logically originated from knowledge management foundations. The term "cognitive computing" suggests computers possessing minds. However, this metaphor misleads—computers, regardless of sophistication, don't think. They support human cognition by reducing cognitive loads through information processing assistance and decision support.
Effective systems surface information as needed. Delivering appropriate information to correct individuals at optimal moments has driven knowledge management for years and remains central to personalization and recommendation algorithms. Organizations fundamentally operate on knowledge flows, consuming information inputs while producing knowledge outputs. Products represent materials combined with knowledge—sophisticated arrangements of basic elements.
Contemporary products across industries demonstrate increased knowledge intensity using less physical material than counterparts manufactured recently. The trajectory toward knowledge-intensive production continues accelerating across sectors.
Technology Limitations Without Foundations
For years, advocacy for knowledge structures supporting artificial intelligence emphasized intentional information architecture approaches. The premise remains straightforward: artificial intelligence cannot exist without information architecture. This conclusion emerged from extensive research examining supposedly AI-powered applications, particularly cognitive technologies like bots and virtual assistants marketed as call center and customer service automation solutions.
Vendor discussions about tool operations, training methods, and functionality development revealed troubling patterns. Responses ranged from proprietary claims to jargon-filled explanations about algorithms learning from data without human intervention. One vendor boldly claimed systems didn't require problem definition. Administrative interfaces displayed question-answer pairs with phrase variations and misspellings—an unmanageable approach to intent classification. Another common non-answer assumed knowledge base existence without addressing the actual challenge: creating knowledge sources.
Artificial intelligence fundamentally performs classification. Algorithms classify signals separating them from noise. Image recognition distinguishes cats from dogs. Medical imaging classifies cancerous versus healthy tissue. Manufacturing quality control identifies defective parts. Cognitive assistants classify phrase variations as intents, then use signals retrieving matching information. Content gets classified as appropriate responses to signals.
However, AI cannot judge content value or fix knowledge base deficiencies. Missing information cannot be algorithmically invented. AI helps improve and curate content through semi-automated indexing, but architecture and reference data must exist first. These include organizationally important terms and concepts. Problem solutions reside in taxonomies and ontologies forming enterprise knowledge scaffolding.
AI demands structure—understanding business essence captured in ontologies containing problems, solutions, roles, processes, questions, answers, topics, content types, customer categories, product classifications, equipment types, attributes, regions, skill areas, research domains. Every operational concept requires definition and mapping for AI functionality.
Knowledge Creation Challenges
Cognitive assistants answering questions and supporting specific tasks require information sources. Training virtual assistants demands identical materials required for human training. FAQ bots need frequently asked questions. Troubleshooting bots require trouble codes and procedures. Knowledge creation remains uniquely human, emerging from creative application of experience and expertise to problems.
Engineers developing product designs must define features, functions, installation procedures, guides, and support content enabling customer, field service, and call center representative success. Substantial captured, codified knowledge exists throughout enterprises: processes, procedures, workflows, system designs, software, reference materials, training content, presentations, methodologies, templates, exemplars, and high-value knowledge assets comprising competitive advantage.
Enterprises compete through knowledge encompassing job-specific expertise acquired over years, customer need understanding, communication approaches, and audience resonance patterns. Over time, organizations learn optimal service delivery, competitive differentiation, customer guidance through selection, usage optimization, troubleshooting, maintenance, upgrades, and replacements. Enormous knowledge bodies span procurement through manufacturing efficiency to transportation logistics. Physical supply chains link inextricably to knowledge and information supply chains.
Product, service, and solution discovery historically relied on direct interpersonal communication or written materials. Digital environments changed this dramatically. Most people research products and options thoroughly before human interaction. Human roles evolved beyond basic education provision formerly central to sales and support.
Looming Expertise Crisis
Corporate environments face escalating knowledge crises as human expertise becomes scarcer. Expertise typically requires years of on-the-job experience mastering domains. Contemporary workers demonstrate less patience for lengthy skill development periods. Work nature shifts fundamentally. Fewer individuals enter workforces planning decade-long industry commitments.
Fortunately, increasing human expertise embeds into applications and products simplifying service and operation. However, successful implementation requires making subject matter expertise explicit through experience capture before retirement or departure. More services enable through digital channels because technologies, unlike human expertise, scale effectively. Again, knowledge sources prove critical—information requires capture and structuring enabling organizational digital machinery serving it contextually for user objective accomplishment.
Large service manuals require chunking providing direct access to specific information answering questions. Years ago, medical insurance claims processors struggled locating needed information buried in three-hundred-page policy documents. People seeking answers don't want searching hundreds of results opening lengthy documents. They want direct answers. Breaking content into pieces benefits humans. Identical structures enabling human question answering enable bot question answering. Well-organized, searchable information delivers this beauty.
Chunked knowledge additionally supports personalization. Messaging pieces recombine for different audiences with variations. Machine learning algorithms further fine-tune variants for particular audiences and contexts. One large technology organization serves four million knowledge objects daily across channels, sites, and contexts. These components enable functionality throughout marketing, service, support, ecommerce, product development, and internal processes supporting those experiences. At such scale, user signals enable automated processes fine-tuning exact answers. Industry, role, equipment ownership, configurations, technical expertise, and weak signals correlate with knowledge usage helping prioritize appropriate information for similar circumstances.
Purpose-Driven Content Development
Recently, an enterprise announced creating groups for AI-ready content. While objectives prove worthy, questions emerge about rationale. Why new groups? Why new content? Content for bots and virtual assistants requires careful alignment with purposes and use cases. It must specifically serve purposes helping users with tasks.
However, shouldn't all content serve purposes with users in mind? Work with a government agency revealed twenty content writers covering particular Medicare areas. Holding up documents and asking about purpose, audience, value, and rationale produced no answers. No one explained group function from customer perspectives. This exemplifies why focused content approaches matter—not AI content groups but purposeful content groups.
Content for AI remains simply content. Mature content operations prove necessary building virtual assistant training materials. Identical discipline makes information easier for everyone while solving immediate problems and preparing for futures. Organizations treating content development purposefully regardless of delivery mechanism position themselves advantageously.
The competitive landscape increasingly rewards organizations recognizing knowledge management as strategic capability rather than administrative overhead. Those systematically structuring knowledge assets for intelligent access extract disproportionate value from AI investments. Those neglecting knowledge foundations experience disappointing technology returns regardless of algorithmic sophistication or vendor capabilities.
Investment priorities prove straightforward. Build comprehensive ontologies capturing organizational knowledge structures. Develop content operations producing purposeful, well-structured information. Establish governance maintaining knowledge quality. Train workers understanding knowledge management principles. Measure value through usage patterns and outcome improvements. These capabilities compound over time, creating sustainable advantages competitors cannot easily replicate.
This article was originally published on CustomerThink and has been revised for Earley.com.