Knowledge Base Design – Best Practices

Last Updated January 7, 2026

Unified Knowledge Base Design for WalkMe DAP, WalkMe Learning Arc, and WalkMe AI

1. Design Objective

Design a single, governed knowledge system that:

  • Serves as the authoritative source of truth
  • Drives consistent guidance across in-app execution, structured learning, and AI
  • Scales across transformations (SAP S/4HANA, Salesforce, ServiceNow, AI)

The knowledge base is process-centric, not content-centric.

2. Logical Architecture Overview

  • Core principle: Knowledge is designed once, then rendered differently by DAP, Learning Arc, and AI.

3. Layer 1: Business & Process Layer

This layer defines what is true, regardless of tools.

Components
  • Business outcomes (KPIs, value drivers)

  • End-to-end processes

  • Roles and responsibilities

  • Policies and compliance rules

Ownership
  • Process Owners

  • Functional Leaders

  • Transformation Office

4. Layer 2: Knowledge Object Layer

This is the heart of the design. All knowledge is stored and governed as structured objects.

Core knowledge objects
Object Purpose Key Attributes
Business Outcome Why the process exists KPI, value driver
Process End-to-end flow Scope, triggers, systems
Role Who performs work Responsibilities
Task Atomic unit of work Inputs, outputs, success
Decision Conditional logic Rules, exceptions
System Context Where work occurs App, screen, transaction
Policy / Control Governance Compliance impact

These objects are tool-agnostic and AI-readable.

5. Layer 3: Delivery & Experience Mapping

Each delivery surface consumes the same objects, but at different depths.

WalkMe DAP
  • Primary object: Task

  • Uses:

    • System context

    • Step-level execution guidance

    • Validation and error prevention

Design rule: DAP never introduces new process logic—only executes approved tasks.

WalkMe Learning Arc
  • Primary objects: Role, Process, Outcome

  • Uses:

    • Role-based journeys

    • Process overviews

    • Curated task enablement

Design rule: Learning Arc reflects the official process taxonomy.

WalkMe AI
  • Primary objects: Task, Decision, Policy

  • Uses:

    • Structured knowledge

    • Governed terminology

    • Context from DAP and Learning Arc

Design rule: AI answers must map back to a known object.

Business Outcome
└── Process
├── Role
│ └── Task
│ ├── Execution Guidance (DAP)
│ ├── Learning Content (Learning Arc)
│ └── AI Answers
└── Policies / Decisions

This hierarchy enables:

  • Traceability

  • Governance

  • Consistent analytics

7. Learning Arc as the Knowledge Index

Best practice

Use Learning Arc as the visible index of the knowledge base.

Learning Arc should:

  • Mirror the process hierarchy

  • Expose role-based entry points

  • Link directly to DAP execution and AI assistance

DAP and AI should feel embedded, not separate systems.

8. Naming & Taxonomy Standards

Required standards
  • One process name globally

  • One task name per action

  • Business language first, system language second

Example:

  • Correct: “Create Sales Order”

  • Avoid: “VA01 Entry”

9. Governance Model

Minimum governance roles
  • Knowledge Owner (Process)

  • Enablement Translator (WalkMe)

  • AI Steward

  • WalkMe Platform Admin

Lifecycle controls
  • Versioning aligned to system releases

  • Review at Fit-to-Standard, Go-Live, and Releases

  • Retirement of obsolete knowledge objects

10. Measurement & Feedback Loop

Measure knowledge effectiveness, not content volume.

  • Task success rate

  • Time-to-proficiency

  • Error reduction

  • AI answer confidence

  • Support ticket deflection

Use analytics to refine the knowledge base—not to add more content.

Was this article helpful?

Thanks for your feedback!

Be part of something bigger.

Engage with peers, ask questions, share ideas

Ask the Community
×