Answer Set Programming (ASP) is a well-known symbolic AI formalism developed in the area of knowledge representation and reasoning. This paper reports on some blendings of ASP with neural approaches, that can be classified as neurosymbolic AI systems. In particular, we describe (i) an industrial application of ASP and neural network that addresses the compliance-checking of electrical control panels, (ii) two possible combinations of ASP with large language models for enabling reasoning from text, and (iii) the usage of machine learning techniques to boost the evaluation of ASP programs by implementing solver selection pipelines.