Getting answers to bullets (numbered items) from text via NLP

This is related to information extraction. In real world data, documents are written in bullets/numbered items form. For example,

How to create a website:

- Get A DNS
- Get a Hosting
- Deploy wordpress or some site ...


above is sample of a structured data. Take another example where content is semi structured,

While sandeep was going to home there was a road on the way he saw a

- Car
- 2 wheeler
- cart

and he carefully crossed the road


If I have to find out the "steps to create a website" or "what is on the road" .. is there an established method using NLP? As some data is semi structured, so simple classification may not work in this case.
Maybe deep learning network or some pretrained model? I have seen google is able to return such results when we search something like "what are steps create a website".

If we want to search for What are the steps to create a website, all you need to do is the find which set of bulleted points is going to answer that. One thing that you can try is that you can vectorize your question using some pretrained model (w2v, GloVe etc) and also encode the descriptions before the bulleted list ( For eg : How to create a website). Once you have the query vector $$q$$ and the descriptions vectors $$d1, d2, d3,..., dn$$ , you can perform cosine similarity of the query vector with all the description vectors and choose the description with the maximum similarity. Following this, the corresponding bullet points will be the answer.