It's an interesting idea, and implementing it in the form of a chat bot has a great social appeal as well: an anonymous friendly comment from "Community" may go down better in some situations than a community member's "What have you tried?" "We're not here to write code for you", etc.
However, it's going to be useful only in a very small fraction of possible scenarios - it would never work in the case you present for example, as it's perfectly possible to ask a good question that happens to contain the sentence "is it possible to XYZ?". And if there's one thing we do not want is a chatbot pestering the askers of perfectly intelligent questions just because some algorithm thinks it might be a bad question.
I think in the end, any automated response would have to be triggered by the community that recognizes the flaw in a question and knows how to address it. We've been down that road several times before (e.g. with "What Stack Overflow is Not"), always unsuccessfully for social reasons (because whatever automated feedback solution you give the community, it tends to get overused or misused over time).
But perhaps there are clear patterns where this might make sense - we should remain open towards thinking about them. I could think of one in the HTML and CSS tags: when there's only text and a link to an external site and no code, point to https://meta.stackexchange.com/questions/125997/ - just as a friendly reminder to make sure the user is aware of the principle.