KM Glossary & Disambiguation (full edition, with Serviceaide notes)

A practical, reader-first glossary you can ship with your KM journey. Each entry is short, action-oriented, and—in key places—annotated with Serviceaide notes so your team sees how the ideas show up in that stack.

Core KM concepts

Knowledge Management (KM)
An operating loop—capture → structure → deliver → use → learn—that makes the best available guidance show up where work happens, with visible ownership and review dates.

Knowledge Object / Answer
A small, reusable unit (how-to, troubleshooting, policy Q&A, decision aid, known error, standard work) with a purpose (“use this when…”), triggers (“you’ll see…”), steps/branches, warnings, owner, and next review.

Information Architecture (IA)
The shape of your knowledge: templates, hierarchy, facets, navigation. IA is the rails; AI is the train.

Knowledge Graph
A model that connects answers to real entities (products/versions, services/CI, policies/dates, roles/permissions) so delivery can pick the right variant automatically and reviews can trigger from change events.

Taxonomy
A shallow, stable classification (e.g., Domain → Topic → Subtopic) for navigation. Keep it light; push nuance into facets.

Facets / Metadata
Retrieval fields (audience, product/version, role, locale, risk) that filter and route answers across surfaces without copy-paste drift.

Tags
Search helpers in user language (nicknames, error codes, acronyms, common misspellings). Tags serve searchers, not authors.

Synonyms
Mappings between how users ask and how authors write (e.g., “token expired” ↔ “authentication timeout”). Maintain via a weekly clinic.

Provenance
Visible trust signals: owner, last reviewed, next review, source link, and change log—shown anywhere an answer appears.

Lifecycle
Rules that keep answers current: time-based reviews plus event-based reviews triggered by releases, policy updates, or change records; and clean retirement/redirects.

Systems & comparisons

KMS (Knowledge Management System)
A governed answer engine with lifecycle and analytics. Publish one canonical answer to many surfaces (agent, portal, VA, in-app) and measure outcomes (assisted resolutions, deflection, MTTR, etc.).
Serviceaide note: Luma Knowledge is the knowledge layer in the Serviceaide ecosystem—centralizing service content and exposing it across channels alongside Luma AI.

KBS (Knowledge-Based System)
A reasoning/decision system that applies rules or cases (eligibility calculators, configuration advisors). Often returns a recommendation and links to the KMS explanation.

KWS (Knowledge Work System)
Workspace for creating knowledge (docs, wikis, design tools, notebooks). Great for drafting and collaboration; not a substitute for a governed KMS.

CMS (Content Management System)
Publishes rich articles/media and webpages. Pair it with your KMS when you need answer lifecycle, provenance, and outcome analytics—not just publishing.

CRM (Customer Relationship Management)
Manages customer data and interactions. Integrates with KMS to surface governed answers inside cases/chats.

LMS (Learning Management System)
Delivers courses and assessments. Links back to KMS for “just-in-time” answer versions of lessons.

SKMS (Service Knowledge Management System – ITIL)
ITIL’s umbrella spanning CMDB, known error DB, and knowledge. Your KMS supplies the governed answers within this fabric.
Serviceaide note: ChangeGear (ITSM) + Luma VA + Luma Knowledge together form an SKMS-style fabric: CMDB/process in ChangeGear, governed articles in Luma Knowledge, conversational delivery/actions via Luma.

Service Catalog (Conversational / VA)
An action layer for approved tasks (“reset MFA,” “unlock account,” “request access”). Good assistants ground replies in KMS content and then execute allowed actions.
Serviceaide note: Luma AI fronts a conversational service catalog—retrieving answers and performing policy-approved actions in IT and business service flows.

Retrieval, conversation & AI (with guardrails)

Semantic Search
Understands meaning (not just keywords) so messy queries land on better results. Measured by top-3 click-through and fewer “no result” searches.

RAG (Retrieval-Augmented Generation)
Chat answers grounded in approved KMS content, with citations and visible trust stamps.
Serviceaide note: Luma’s “knowledge-centered” narrative—ground conversations in enterprise knowledge, then act—maps to this pattern.

Knowledge-Centered Virtual Agent (VA)
An assistant that cites governed answers, asks clarifying questions when confidence is low, performs only policy-approved actions, and escalates with context.
Serviceaide note: Luma Virtual Agent is positioned explicitly as knowledge-centered from v2.x through Luma 3.0.

Hallucination
Confident, unsourced output. Prevent by grounding on the KMS, showing provenance, constraining scope, and refusing when no source qualifies.

Prompt / Model Governance
Version prompts/models that pass your golden-set tests; log sources and completions; run periodic safety/“prompt-injection” checks.

Generative Drafting (Author Assist)
Model creates a first draft inside your template from long inputs (PIRs, policy PDFs, release notes). Human owners edit, approve, and publish.

Serviceaide note for this section: Luma 3.0 messaging highlights “AI-everywhere” across IT/ESM—intents, knowledge retrieval, and ticket actions living in one conversational flow. Keep KMS governance visible in every surface.

Processes, models & cycles

KM Cycle / Value Chain
Different labels, same loop: capture (signals) → structure (templates, metadata) → deliver (multi-surface) → use (solve/act) → learn (analytics → improvements).

Known Error (ITIL)
Documented root cause/workaround awaiting permanent fix. Short review cadence; retire when the fix ships.

Standard Change (ITIL)
Pre-approved, low-risk change with a repeatable procedure—prime KMS material.

Decision Tree / Decision Aid
Guidance that adapts based on answers. Must show preconditions, branches, and escalation rules; perfect for KBS+KMS combos.

Policy-to-Practice
A short explainer that turns policy text into operational guidance with effective dates and real examples.

Roles, ownership & governance

Executive Sponsor
Funds the program; expects outcomes (assisted resolution, deflection, MTTR) not page counts.

KM Program Lead
Owns standards, lifecycle, analytics, and enablement; keeps the flywheel turning.

Domain Owner
Accountable for accuracy within a scope (HR, IT, Field, CX). Exactly one named owner per article.

Approver
Signs off per risk tier: standard (domain approver) vs regulated (domain + compliance/legal).

Contributor
Anyone close to the work who proposes edits or drafts; on-ramps must be frictionless (from tickets, chats, docs).

Findability toolkit

Title & Snippet
Your cheapest boost. Front-load trigger phrases and a “use this when…” line so users pick the right result.

Synonym Clinic
Weekly 15-minute pass over failed queries and high-bounce intents to add mappings that match user language.

“Used then Escalated”
The answer helped but didn’t finish the job—top signal for backlog and edits.

Pogo-Sticking
Users click and bounce—often a title/first-screen mismatch or an encyclopedia page that needs splitting.

Templates (the six that cover ~90%)

How-To (task), Troubleshooting (symptom → fix), Decision Tree (branched), Policy Q&A (eligibility, steps, dates), Known Error (workaround, fix path), Standard Work (visual steps, safety callouts).

Metrics that matter

Article-Assisted Resolution (AAR)
An answer contributed to solving a case. Track by topic/channel.

Self-Service Solves
Sessions that end without escalation after a knowledge click.

MTTR / FCR / AHT
Incident repair time, first-contact resolution, average handle time—watch deltas where you shipped answers.

Change Failure Rate
Falls when pre-checks/rollback steps live as knowledge in change plans.

Exception Rate (HR/Policy)
Drops after policy-to-practice explainers publish with clear examples/effective dates.

Search Health
Top-3 CTR on key intents; shrink “no result” queries over time.

Industry adapters (quick cues)

HR / People Ops
Policy Q&A + decision aids; portal + case surfaces; legal approval; success = deflection and exception rate down.

Healthcare Administration
Decision aids tied to payer bulletins; show effective dates and source links; keep snapshots for audit.

IT / ITIL
Agent-assist cards keyed to CI/error strings; retire known errors post-fix; success = MTTR ↓, AAR ↑.

Field Service / Manufacturing
Visual standard work; model/firmware facets; QR to variant; offline first; success = first-time fix ↑, repeat rolls ↓.

Customer Support / CX
One canonical answer with audience variants; VA grounded in KMS; success = self-service solves ↑, FCR ↑.

Partner Enablement
Audience-gated answers, no forks; success = time-to-enable ↓, partner case volume ↓.

Serviceaide note for this block: Luma VA runs across IT and business service domains; Luma Knowledge provides the governed content that powers agent cards, portals, and conversational replies.

Personal & team knowledge

PKM (Personal Knowledge Management)
Fast, searchable capture for individuals (notes, snippets, screenshots). Bridge to team plays; promote repeatable items into the KMS.

Team Plays
Reality-first guides for “how we do this here.” Mark as draft, link to the canonical KMS answer when promoted, and stop editing the draft.

Promotion
Three triggers: frequency (used 3+ times/week), risk (policy/safety/money), portability (useful outside the team). Paste into the right KMS template, set owner/audience/review date, approve, and link back.

Disambiguation (non-KM meanings you’ll bump into)

km / km/h
Kilometer (distance) and kilometers per hour (speed). Not knowledge management.

Km (Michaelis–Menten)
Enzyme kinetics constant. Different universe.

KMS (Key Management Service, cryptography)
Encryption key management (e.g., AWS KMS). Not a Knowledge Management System.

“5 C’s / 4 C’s / 5 P’s / pillars”
Vary by field. In KM they’re teaching mnemonics (e.g., Culture, Content, Context, Community, Change). Don’t treat them as standards.

PMBOK Knowledge Areas
Project management concept; adjacent when mapping change and governance, not KM itself.

IA (outside KM)
Sometimes means “Intelligent Automation.” In this glossary, IA = Information Architecture.

Cross-references (route readers to the right deep dives)

• Structure & findability: Page 1 (Taxonomy, Tags, Synonyms, Relevance)
  
  
• Owners, reviews, audits: Page 2 (Governance, Audit & Compliance)
  
  
• By-function playbooks: Page 3 (Industry & Function)
  
  
• What examples look like: Page 4 (KMS / KBS / KWS Examples)
  
  
• AI that stays grounded: Page 6 (AI & IA in KM)
  
  
• Personal → Enterprise handoff: Page 8 (PKM & Team KM)
  
  

Extra Serviceaide tie-ins (quick anchors)

• Luma Knowledge (KMS layer): centralizes service knowledge and exposes it to VA/portal/agent.
  
  
• Luma Virtual Agent (knowledge-centered): understands intents, retrieves governed answers, and can take actions (ticketing, updates, approvals) within policy.
  
  
• Luma 3.0 (AI-everywhere): generative features layered across service workflows; keep provenance visible and refusal rules intact so speed doesn’t outrun trust.
  
  
• ChangeGear + Luma: SKMS-style fabric for ITIL—CMDB/process + knowledge + conversational action.
  
  
• Field & mobile cues: VA entry points (chat, voice, mobile) plus QR/context help—good match for launching model/firmware-specific procedures.
  
  

A last word on usage

When in doubt, draft the first line of any entry (or answer) as “Use this when…”, front-load the trigger phrase users actually type, keep steps atomic, put warnings first, and show owner / last reviewed / next review. That pattern turns pages into answers—and answers into outcomes—no matter which platform you’re on.

‍