Who uses ontologies in practice and for what use case?

They're increasingly being used in the building automation industry to describe HVAC and lighting systems for analytics purposes.

RDF suits because inference allows it to deal with massive scope of data (the amount of different hvac products and configurations in buildings is crazy), and prevents vendor lock in (a surprisingly common concern in that industry).

The two relevant schemas are Project Haystack[0], and Brick Schema[1].

As far as I know, the most mature analytics provider is Skyspark[2].

Super interesting space.

[0] https://project-haystack.org/ [1] https://brickschema.org/ [2] https://skyfoundry.com/product

ISA² [0] (Interoperability solutions for public administrations, businesses and citizens) uses various ontologies, like Core Vocabularies [1].

[0] https://ec.europa.eu/isa2/solutions_en

[1] https://ec.europa.eu/isa2/solutions/core-vocabularies_en

(but maybe you meant this question in a more general sense, not specific examples)

I find ontologies useful in the preamble of requirements documents to define the terms and concepts that will be later used in the body of the document.

Generally they are simple and accepted concepts for the domain specialists, but not well understood by programmers who have to read the requirements and business analysis documents.

Healthcare. Biomedical informatics IIRC did/does a lot with this in nosology.