An application ontology fits knowledge into a machine-readable format for a specific purpose, such as running a specialised AI agent. Unlike a domain ontology, which provides the map and rulebook for a single field of expertise, an app ontology can be customised to span several domains, or focus on one area of interest within a domain.
While some app ontologies introduce new concepts and designs, ontology standards tend to be used where possible, enabling the app to share data smoothly with other systems (interoperability).
What is an ontology?
Ontologies frame knowledge using controlled vocabulary in a way that also sets out how those words connect to other concepts in the space. In the field of cardiology, for example, an ontology might define the technical definition of ‘heart attack’ and ‘diagnosis’, but also clarify how they connect to ‘treatment’. This enables AI to draw logical conclusions from the information (inference).
What are the benefits of an app ontology?
When information is structured and governed by an app ontology, AI services become more domain intelligent, meaning their outputs are more useful and trustworthy. The AI’s outputs are more trustworthy because they are governed and validated by the organisation’s rules and constraints. They are more useful because humans can interact with the AI agent naturally via an optimised large language model (LLM), and check its reasoning (XAI) for compliance.
A recent study published in the Journal of Engineering Design showed the benefits of tailoring an engineering ontology on top of a high-level ontology for a specific applied domain.
Why are app ontologies quicker to develop?
Complex domain ontologies can take years for experts to develop, requiring consensus-building approaches such as the Delphi Method. An app ontology can avoid replicating this time-consuming work by importing relevant terms from the existing domain ontology, then potentially customising new ones for added granularity.
Modern ontology-as-a-service (OaaS) platforms can be used to accelerate the development and implementation process further.
Is it worth creating an app ontology?
Re-using an existing ontology standard should always be the default choice of an app ontology developer, firstly because it’s easier, but also because it will ensure that key terms align with other systems. This will help the app to scale.
Over time however, new ontological concepts and subsets are always needed, reflecting changes or innovations in the domain. While modern ontology software gives non-technical people more tools to build an ontology on their own, an experienced ontology engineer might still be needed for more complex or ambitious apps.