Knowledge Modeling Analysis method

Competency Questions

Defines answerable domain questions as testable requirements for an ontology.

Competency questions state in natural language which inferences an ontology should enable. Terms, quantification, example answer, and test query are explicit.

Core question

Which domain questions must the ontology answer with defined data and inferences?

Plan Competency Questions Method guide

Start comparison

Give feedbackIs this method useful to you?

Participants: 2-8
Time: 2-4 h

Competency Questions uses translate a natural-language domain question into terms, formal query, test data, and expected answer. Assessments remain justified and context-bound.

Paper illustration of Competency Questions with its method-specific working model.

Elicit question
Elicit question
Collect real decision questions with domain actors.
Specify
Specify
Clarify quantifier, horizon, terms, and desired answer.
Formalise
Formalise
Translate CQ into a query and required axioms.
CQ tests
CQ tests
Run positive, negative, and boundary test data.

Update after new evidence or a material context change and retain the previous version.

A good fit when

Before and during ontology development to clarify scope and acceptance.

Ideal for

Ontology Engineering
Knowledge Graph Scoping
GraphRAG Pipelines
Schema Design

Not good for

Rein explorative Domain Discovery ohne Zielsystem
Reine Datenintegration ohne semantisches Modell

Outputs

competency question test suite
competency question test suite with traceable sources, assumptions, and domain decisions.
Decision and review trail
Owners, dates, and next review for competency question test suite.

competency question test suite

Fictional worked example

translate a natural-language domain question into terms, formal query, test data, and expected answer as a precise working structure with a fictional example.

Competency question

User · question · decision

Which open actions are overdue?

Semantics

Terms · quantifier · horizon

status=open; dueDate<today.

Test case

Data · expected answer · negative case

A-17; completed A-18 excluded.

Formalisation

Query · axioms · version

SPARQL + hasOwner/memberOf.

The example demonstrates the domain logic of Competency Questions; worksheets remain blank.

Grüninger and Fox: Methodology for the Design and Evaluation of Ontologies

First step

Start with the core question: Which domain questions must the ontology answer with defined data and inferences?

Open the method guide

Good to do first

Knowledge Modeling

Domain Storytelling

Supplies narrated domain knowledge from which answerable competency questions are derived.

Where to go next?

Knowledge Modeling

Ontology Design Patterns

Uses the test questions to select fitting design patterns for the ontology.

Paper illustration of a Concept Map with concepts and labelled links

Knowledge Modeling

Concept Mapping

Translates answered questions into explicit concepts and relationships.

Deep dive

Purpose

Competency Questions translate a domain model into concrete questions that it must be able to answer. The method keeps the model's scope clean and prevents pretty but useless structures.

How it works

First, real questions from everyday domain work are collected. The team then checks whether a planned model can answer these questions or whether concepts and relationships are missing. This makes domain usefulness the actual quality criterion.

In detail

Competency Questions are effective because they measure a model against the questions that actually need to be answerable in everyday domain work. The method keeps modeling honest as soon as every question is traced back to concepts, relationships, and necessary distinctions. Good use takes care to include operational examples, edge cases, and later query forms, so that no elegant schema emerges that misses practical information needs.

Facilitation

At the start, collect real questions from reports, search cases, or decision routines and sort them by domain relevance. The lever lies in questions that allow multiple interpretations; there the meaning must be sharpened before the model grows further. End with a small set of prioritized questions against which new model parts will be checked in the future.

The method guide walks execution through 4 phases, timing, 4 pitfalls, and clear stop criteria.

Open method guide

Your perspective

Does this method help with your work?

Your feedback helps shape the next improvements to methodatlas.

Sources and topics

Topics

Ontology Knowledge graph Scope Requirements Semantic

Sources

Grüninger and Fox: Methodology for the Design and Evaluation of Ontologieseil.utoronto.caPaper by Michael Gruninger and Mark Fox (IJCAI 1995, University of Toronto); introduces competency questions for evaluating ontologies.

In playbooks

Prepare Knowledge ModelingMove from a knowledge domain to questions, concept model, schema, and validation rules.

Artifact templates

Competency Questions: WorksheetPlan Competency Questions around its domain steps and the concrete decision question.

markdown

competency-questions-working-template.md

Plan Competency Questions around its domain steps and the concrete decision question.

Competency Questions: Worksheet

Question

To be confirmed

Desired outcome

To be confirmed

Scope

Full workshop Time: 60–90 minutes initially; ongoing test maintenance

Complete this worksheet on paper or in your own document during the work.

Preparation for this scope

Method setup

Materials: Use cases and decision questions, domain vocabulary, representative instance data, shared workspace, source links, and decision log.
Roles: Domain expert · ontology engineer · data consumer
Advance information: Collect Use cases and decision questions, domain vocabulary, representative instance data in advance and mark open assumptions.
Overall time needed: 60–90 minutes initially; ongoing test maintenance
Setup: Prepare translate a natural-language domain question into terms, formal query, test data, and expected answer visibly and calibrate assessment terms before starting.