Secured facts through intelligent knowledge modelling.

Knowledge Graphen & Retrieval Augmented Generation

The perfect



With our Knowledge Graph technology, we connect your entire company knowledge and prepare it for dialogue management in natural language. This gives you complete control over the responses of your virtual assistants. The combination of RAG technology and knowledge graphs enables you to generate answers based on the content provided and avoid or minimise the risk of hallucinations or incorrect answers.

We would be happy to show you the advantages of our knowledge graphs in a short online demo. Simply arrange a free consultation.

Features and functions


Knowledge Graph

A knowledge graph stands for a special kind of knowledge representation.

Data/content is modeled in order to describe relationships between individual “entities”, make them machine-readable and to enable answering complex queries.


Retrieval Augmented Generation

In the Onlim RAG approach, a knowledge graph is used to provide facts in the form of structured information about entities and their relationships for answering.

When a query is made, the system accesses the relevant context from the knowledge graph and uses it to generate more accurate and fact-based answers.


Data modelling

In knowledge graphs, facts are stored in the form of edges between nodes in a graph/network.

In addition, the schema of the data is also stored in the graph (e.g. class hierarchies) and semantic models are developed. This adds semantic understanding to machine readability.


Implicit knowledge

New facts can be derived from existing knowledge by linking the individual entities.

Example: Falco was a musician. He lived in Vienna, Vienna is a city in Austria. The question about Austrian musicians must therefore also result in Falco, without “Falco lived in Austria” having been modelled directly in the background.


Conversation Optimization

The Knowledge Graph in combination with RAG offers several advantages for conversation optimisation:

Expansion of language understanding

The knowledge stored in the Knowledge Graph, e.g. names and synonyms of entities, is used to improve the bot’s language understanding.

Faktenbasierte Antworten

The combination of RAG technology with knowledge graphs makes it possible to generate answers based on the content provided and to avoid or minimise the risk of hallucinations or incorrect answers.

Improved answers and dialogues

The use of a knowledge graph and the underlying data modeling make it possible to answer very specific and complex questions.

Calculations directly in chatbot

By modeling the corresponding data in the background, it is possible to perform calculations directly in the chatbot, e.g. prices can be calculated.

Reasoning / generation
of new knowledge

Existing knowledge is extended by connecting to other nodes, new relationships are thus created or generated.

Conversational AI on a new scale.

Request a non-binding demo now.