The term "compositional AI" originates from the fields of artificial intelligence, digital transformation and automation. Compositional AI describes a method in which various small, specialised AI models are combined with each other to solve complex tasks. Unlike conventional "monolithic" AI systems that want to do everything at once, with compositional AI many small units share the work.
Imagine a modular system: There is a separate, specialised AI module for each subtask. If these are skilfully combined, the overall system can handle tasks that would be too difficult for a single AI model.
A practical example: In automated customer service, individual AIs are used - one to understand the language, another to analyse the mood in the text and another to select suitable answers. The combination of these specialised "building blocks" results in a flexible and particularly powerful system that can be easily adapted to different requirements.
Compositional AI thus facilitates the development of smart, versatile applications and makes complex technologies more accessible to companies of all sizes.