Exercise 1: AI Use-Case and Knowledge Context Canvas
Purpose: Clarify the business purpose and information context before designing IA artifacts.
Participant work:
Describe the AI solution and intended business outcome.
Identify primary users, tasks, decisions, and questions the solution must support.
Define required knowledge, scope boundaries, assumptions, exclusions, and constraints.
Record privacy, security, regulatory, records, and responsible AI considerations.
Required output:Â Blueprint Section 1: AI use case, users, decisions, knowledge needs, boundaries, assumptions, and constraints.
Exercise 2: Content and Knowledge Source Inventory
Purpose: Identify and prioritize the repositories, applications, records, and expert knowledge required by the AI use case.
Participant work:
List priority sources and systems of record.
Capture source owner, format, authority, access, sensitivity, freshness, lifecycle, and usage rights.
Identify duplicated, obsolete, conflicting, restricted, or unsupported sources.
Map each source to the tasks or decisions it supports.
Required output:Â Blueprint Section 2: Prioritized source inventory and ownership map.
Exercise 3: AI-Readiness and Risk Assessment
Purpose: Determine whether priority sources are suitable for AI ingestion, retrieval, grounding, and citation.
Participant work:
Score accuracy, authority, completeness, currency, structure, metadata quality, provenance, searchability, and retrieval suitability.
Assess permissions, sensitivity, retention, external-use restrictions, and AI approval status.
Capture evidence for each score rather than relying on unsupported judgment.
Prioritize remediation actions, dependencies, and unresolved questions.
Required output:Â Blueprint Section 3: Readiness findings, material gaps, risk priorities, and remediation actions.
Exercise 4: AI-Ready Content Model
Purpose: Create a conceptual model that gives the AI solution clear content types, components, relationships, identifiers, and lifecycle states.
Participant work:
Identify priority content types and business objects.
Define reusable components such as title, summary, purpose, audience, procedure steps, source, and related content.
Model relationships among policies, procedures, standards, FAQs, products, services, decisions, and experts.
Document lifecycle, versioning, reuse, citation, and chunking implications.
Required output:Â Blueprint Section 4: Conceptual content model and relationship definitions.
Exercise 5: Metadata and Provenance Schema
Purpose: Define the minimum metadata required for trustworthy discovery, filtering, ranking, access, citation, and governance.
Participant work:
Define field name, definition, purpose, data type, allowed values, source, requirement level, owner, and validation rule.
Include descriptive, structural, administrative, security, rights, provenance, retention, and AI-use metadata.
Specify source authority, approval for AI use, review and expiration dates, audience, jurisdiction, and retrieval priority.
Document how metadata supports access filtering, explainability, citation, and auditability.
Required output:Â Blueprint Section 5: Metadata, classification, provenance, and access-control schema.
Exercise 6: Taxonomy and Semantic Layer
Purpose: Create a minimum viable controlled vocabulary and determine where richer semantic relationships are justified.
Participant work:
Define preferred terms, synonyms, acronyms, non-preferred terms, and scope notes.
Create broader/narrower relationships and relevant facets.
Identify important entity types and relationships.
State whether taxonomy alone is sufficient or whether ontology or knowledge-graph capabilities are needed.
Required output:Â Blueprint Section 6: Taxonomy, synonyms, facets, entities, and semantic-layer recommendations.
Exercise 7: Retrieval and AI Experience Design
Purpose: Translate IA decisions into requirements for search, RAG, copilots, assistants, and agents.
Participant work:
Select retrieval methods and define filtering and ranking rules.
Document chunking assumptions, context assembly, source authority, and citation requirements.
Design facets, answer structure, confidence cues, 'show your source' behavior, escalation, and feedback.
Define source-grounded prompt components and restrictions.
Required output:Â Blueprint Section 7: Retrieval, grounding, citation, prompt context, interaction, and feedback requirements.
Exercise 8: IA Governance and Operating Model
Purpose: Define the accountability and processes required to maintain IA artifacts and approved knowledge sources throughout the AI lifecycle.
Participant work:
Assign owners, stewards, decision rights, reviewers, and escalation authorities.
Define create, review, approve, publish, reuse, change, retire, and archive workflows.
Set governance requirements for content, metadata, taxonomies, knowledge bases, prompt libraries, and retrieval configurations.
Establish review cycles, quality controls, issue management, and change approval.
Required output:Â Blueprint Section 8: Governance roles, RACI, policies, workflows, controls, and review cycles.
Create an IA-related AI risk and control matrix.
Define tests for relevance, groundedness, completeness, citation accuracy, freshness, access leakage, and answer usefulness.
Establish KPIs, thresholds, monitoring cadence, owners, and escalation criteria.
Identify business value measures and user trust indicators.
.png)
.png)
.png)
.png)