AI-Enabled Knowledge Management Training Course

Participants can design core knowledge structures that improve content classification, discovery and retrieval.

• Enterprise taxonomy design using facets, preferred terms, synonyms and relationships
• Metadata schemas align Dublin Core elements with business retrieval requirements
• Controlled vocabularies reduce ambiguity in policy, casework and technical repositories
• Content models define mandatory fields, ownership, review cycles and retention
• Knowledge graph triples connect entities, concepts, documents and authoritative sources
• Findability testing measures search success, query reformulation and time-to-answer

Participants can apply records management controls to information made available through AI knowledge services.

• ISO 15489 principles set authenticity, reliability, integrity and usability requirements
• Retention schedules link record classes to legal, regulatory and operational triggers
• Disposition authorities document approval, holds, destruction and permanent preservation decisions
• Legal holds suspend routine disposal for investigations, disputes and audits
• Information classification labels govern access, handling, sharing and storage controls
• Records inventories expose unmanaged repositories, duplicate content and ownerless data

Participants can assess where generative AI is suitable and unsuitable for knowledge management tasks.

• Large language model architecture explained through tokens, embeddings, transformers and inference
• Retrieval-augmented generation grounds answers in approved organisational knowledge sources
• Prompt patterns specify role, task, context, constraints and output format
• Hallucination, bias and prompt injection assessed through realistic knowledge scenarios
• Human-in-the-loop review assigns accountability for generated summaries and answers
• Use-case selection scores value, feasibility, risk, data readiness and adoption

Participants can specify the content preparation and retrieval design needed for trustworthy AI answers.

• Source selection prioritises authoritative, current, rights-cleared and well-structured content
• Chunking strategies preserve semantic context across policies, procedures and long documents
• Embedding models convert content segments into vectors for semantic retrieval
• Retrieval pipelines combine vector search, keyword search, filters and reranking
• Citation design links generated answers to source passages and version dates
• Knowledge base refresh workflows manage ingestion, approval, deprecation and reindexing

Participants can define governance controls for AI-supported knowledge and records management services.

• NIST AI RMF functions organise govern, map, measure and manage activities
• AI governance roles separate system ownership, content stewardship, risk oversight and approval
• Risk registers record harms, controls, residual risk, owners and review dates
• Privacy impact assessments examine personal data, purpose limitation and data minimisation
• Access controls apply least privilege, role-based permissions and segregation of duties
• Model cards document intended use, limitations, evaluation results and monitoring

Participants can test AI knowledge services and define monitoring evidence for ongoing control.

• Evaluation sets test factuality, citation accuracy, relevance, completeness and refusal quality
• Groundedness metrics compare generated claims against approved source evidence
• Red-team testing probes jailbreaks, data leakage, toxic output and misinformation
• Acceptance thresholds define release criteria, escalation routes and remediation actions
• Sampling plans review live interactions by risk tier and user group
• Audit trails retain prompts, sources, outputs, approvals and system versions

Participants can design AI knowledge services that meet user needs and include clear escalation pathways.

• User journey mapping identifies knowledge needs, search barriers and decision moments
• Service blueprints assign frontstage interactions, backstage workflows and ownership
• Conversational design creates clarifying questions, fallback paths and escalation rules
• Answer templates distinguish policy guidance, procedural steps, summaries and referrals
• Accessibility requirements address plain language, screen readers and multilingual content
• Change management plans address trust, incentives, training, feedback and adoption

Participants can measure service performance, manage AI incidents and assess supplier control arrangements.

• Knowledge performance scorecards track reuse, search success, freshness and contribution rates
• AI service metrics measure containment, resolution quality, latency and escalation
• Benefits realisation maps time saved, error reduction, risk avoidance and service outcomes
• Content health indicators flag stale articles, broken links, duplicates and gaps
• Incident management handles harmful outputs, security events and knowledge corrections
• Vendor due diligence examines data use, hosting, subcontractors and exit provisions

Participants can produce a staged implementation plan with accountable owners, controls and performance measures.

• Use-case portfolio ranks initiatives by value, risk, readiness, cost and dependencies
• Target operating model defines decision rights, teams, processes, platforms and controls
• 90-day implementation plan sequences pilots, content remediation, testing and rollout
• RACI matrix assigns accountable owners for knowledge, records, AI and security
• Procurement requirements specify interoperability, auditability, data residency and service levels
• Personal application plan commits first actions, stakeholders, measures and review dates