https://en.wikipedia.org/wiki/AI_effect
There's this weird effect that's been observed wherever once an AI technique starts working, it becomes a field in its own right and ceases to be considered "AI". So for example, beam search and A* were once considered AI; now they're just algorithms that everyone should know. SVMs, decision trees, linear regression, and other statistical machine-learning techniques were once considered AI; now they form the basic knowledge for the field of data science. Computer vision, handwriting recognition, and OCR were once considered AI (indeed, they were some of the first applications of artificial neural networks); now they are fields in their own right. Ditto machine translation and voice recognition. Even information retrieval with its subfields of synonym recognition, language models, bigram recognition, NLP, named entity recognition, etc. were all considered AI.
That "AI" is considered the hot sector means that VCs don't have a clue what is coming next.
I.e medium has a web development core, but that doesn't put it in the web development sector.
· Sales and marketing integration
· Enterprise search
· Multi-platform marketing
· Slackbot for sales teams
· Predictive analytics on customer lifecycle
A better taxonomy would indicate their sector and either be AI powered <sector> or just include their AI processes in the description.
The thing is though that AI opens lots and lots of new business opportunities and if these companies weren't doing nominally AI processes they wouldn't exist.