Skip to content
KRENContact
AXEmbeddedSI
CubeonQDataQFactoryAgentQOntoQQDriveEVCPQVision
QtTelit CinterionToradexVisualOnTuxeraProtoPie
About OCUBEHistoryUse CasesLocations

Ontology-Driven Knowledge Platform

OntoQ

Ontology-Driven Knowledge Platform

OntoQ

Organise business knowledge once. Ask questions naturally. An ontology platform that brings documents, data, and decision evidence together.
OntoQ business dictionary and natural-language query workspace with sample dataOntoQOntoQ

Overview

Connect business knowledge. Understand the evidence.

Search data and documents using registered business dictionaries and rules, and see why each answer was reached.

Business dictionaries and rules

Map terms in questions to business concepts and apply registered rules consistently.

Unified document, DB, and graph search

Find business data, documents, and relationships through a single question.

Answers with visible evidence

Review applied rules, data values, source documents, and relationship graphs alongside each answer.

Extend to new domains without code

Expand to other teams and workflows with dictionary and rule definitions and data connections.

Comparison

Beyond document search: business data and decision evidence

OntoQ adds dictionaries, rules, business databases, and relationships to conventional RAG document search.
CategoryConventional AI (chatbot)Conventional RAG serverOntoQ
Knowledge sourcesPretrained general knowledgeRegistered documentsDictionaries and rules + business DB + documents + relationships
Facts and aggregatesCannot access company dataRelies on document textQueries business databases directly
Decision criteriaMay vary by question or contextAI interprets document explanationsApplies registered rules consistently
Following relationshipsNot availableLimited to similar documentsTraverses connections and displays a graph
Answer evidenceAbsent or unclearSource documentsSource documents + applied rules + relationship graph
When no results existMay generate a plausible answerAnswers from document contentReports no results without asking AI

  • Knowledge sources

    Conventional AI (chatbot)
    Pretrained general knowledge
    Conventional RAG server
    Registered documents
    OntoQ
    Dictionaries and rules + business DB + documents + relationships

  • Facts and aggregates

    Conventional AI (chatbot)
    Cannot access company data
    Conventional RAG server
    Relies on document text
    OntoQ
    Queries business databases directly

  • Decision criteria

    Conventional AI (chatbot)
    May vary by question or context
    Conventional RAG server
    AI interprets document explanations
    OntoQ
    Applies registered rules consistently

  • Following relationships

    Conventional AI (chatbot)
    Not available
    Conventional RAG server
    Limited to similar documents
    OntoQ
    Traverses connections and displays a graph

  • Answer evidence

    Conventional AI (chatbot)
    Absent or unclear
    Conventional RAG server
    Source documents
    OntoQ
    Source documents + applied rules + relationship graph

  • When no results exist

    Conventional AI (chatbot)
    May generate a plausible answer
    Conventional RAG server
    Answers from document content
    OntoQ
    Reports no results without asking AI

Features

From questions to evidence, in one platform

Users ask questions and review evidence. Administrators manage business knowledge and connections.
Natural Language Q&A

Natural-language questions and answers

Ask questions using familiar business terms. Continue the conversation to explore the results.

Search modes
Automatic (default), structured, semantic, graph, and hybrid
Query modes
Stored results (default) or on-demand calculation
Continue conversations
Review conversation history and each answer’s business domain
Explainable Evidence

Evidence-based answers (XAI)

See why an answer was reached by reviewing the applied rules and sources directly below it.

Evidence cards
Expand applied rules, data values, and source documents
Relationship graph
See the equipment, documents, and entities linked to an answer
Detailed validation
Administrators can also inspect query commands and search plans
Rule-based Reasoning

Rule-based decisions

Register business criteria, such as threshold violations, and apply consistent rules to the same data.

Register decision rules
Define explicit rules for the criteria used in your work
Consistent decisions
Classify entities using registered criteria and show evidence
Update criteria
Revise rules and regenerate knowledge data to apply changes
Domain Management

Business domain management

Organise dictionaries, rules, and data connections by domain. Use the same approach across new teams and workflows.

Register definitions
Register concepts, properties, rules, and data mappings
Validate connections
Check dictionary, DB, and document connections when saving
Switch domains
Activate a domain to apply it from the next question
Document & Graph Search

Document and relationship search

Explore meaning in documents and links between entities to understand context beyond document search.

Semantic search
Search PDF, Word, and Excel content by meaning
Graph search
Follow related entities and documents in a relationship graph
Start with documents
Begin document search without setting up a business domain
Admin & Operations

Administration and operations

Manage documents, knowledge, and AI instructions together. Monitor platform operations within your organisation.

AI instructions
Review answer instructions and restore previous versions
Status monitoring
Check login, storage, AI server connections, and synchronisation
Safe queries
Use read-only access and query result limits

Explainable Evidence

See why decisions were made in a relationship graph

Connect business concepts to real data and review the applied rules and decision results.
OntoQ relationship graph showing equipment, sensors, locations, and F1 high-temperature rule inference
Red dashed lines show equipment classified by the F1 high-temperature rule. The original Korean graph illustrates sample equipment, sensors, and locations.

Runtime Pipeline

How a question becomes an answer backed by evidence

AI plans the search. Validated programs perform the actual queries with read-only access.
  1. 1

    Ask a question

    Use familiar language. The search mode is selected automatically.

  2. 2

    Connect concepts

    Map question terms to concepts and relationships in the business dictionary.

  3. 3

    Plan the search

    AI plans what to find and which search methods to use.

  4. 4

    Build and validate queries

    Programs create query commands and validate them against registered definitions.

  5. 5

    Search in parallel

    Run structured, semantic, and graph search together. Use remaining results if one search fails.

  6. 6

    Answer with evidence

    Build answers only from collected evidence. Provide evidence cards and relationship graphs.

Architecture

An ontology platform that runs entirely on your own servers

Separate question processing from knowledge generation and connect business data with read-only access.

  1. Users (web interface)

    End usersAdministrators

  2. Access

    Gateway

  3. Query processing

    Query processing serverAI model server

  4. Knowledge stores

    Knowledge data storeGraph storeDocument search storeFile store
    Synchronisation processBusiness database
View component details
End users
Ask questions · review evidence
Administrators
Manage documents, domains, and instructions
Gateway
Authentication and permissions
Query processing server
Analyse questions · orchestrate search Compose answers (rag_service)
AI model server
On-premise GPU · vLLM
Knowledge data store
Dictionary and rule data (Fuseki)
Graph store
Relationship data (Neo4j)
Document search store
Semantic search (Qdrant)
File store
Source documents and definitions (MinIO)
Synchronisation process
Definitions + business data → knowledge data
Business database
PostgreSQL Read-only

Users (web interface)

End users

Ask questions · review evidence

Administrators

Manage documents, domains, and instructions

Access

Gateway

Authentication and permissions

Query processing

Query processing server

Analyse questions · orchestrate search Compose answers (rag_service)

AI model server

On-premise GPU · vLLM

Knowledge stores

Knowledge data store

Dictionary and rule data (Fuseki)

Graph store

Relationship data (Neo4j)

Document search store

Semantic search (Qdrant)

File store

Source documents and definitions (MinIO)

Synchronisation process

Definitions + business data → knowledge data

Business database

PostgreSQL Read-only

Question processing flowKnowledge generation flow

Use Cases

Different workflows. The same approach to evidence.

Register domain dictionaries, rules, and data connections to use them on the same platform.
Manufacturing · Equipment

Equipment and process checks

Question
“Find devices in the factory with problems.”
Process
Map dictionary concepts → high-temperature rule → equipment DB query → related documents
Result
Abnormal equipment, locations, applied rules, and sensor evidence
Quality · Compliance

Test, quality, and compliance search

Question
“Why was the test repeated, and what was the procedure?”
Process
Query test records and approvals + semantic search of SOP documents
Result
Reasons, procedures, approvals, and source documents
Operations · Logistics

Operational status queries

Question
“Which deliveries are delayed, and what contract clauses apply?”
Process
Query delivery and inventory status + search contracts and policy documents
Result
Status tables, related clauses, and relationship graphs

Ideal Use Cases

Built for these teams and workflows

For work that combines documents and business data and requires evidence behind decisions.

Teams with business data in databases

Environments that need natural-language queries for facts, status, and aggregates stored in business databases

Work with clear decision criteria

Work that registers business criteria as rules and applies the same standards to the same data

Decisions requiring verifiable evidence

Work that reviews applied rules, data values, and source documents alongside answers

Work combining documents and data

Work that checks business status, related documents, and connections between entities in one place

Organisations expanding across departments

Teams extending use through new domain dictionaries, rule definitions, and data connections

Environments with restricted data sharing

Organisations that operate internal data using AI models and storage on their own servers