Everyone else is focusing on news articles as well, so that's not a problem; however without comparison to other approaches it's impossible to tell if the solution is 'powerful' or even any good at all.

This here - http://nlpprogress.com/english/summarization.html - would be an overview of what algorithms are currently considered 'good' and what results they achieve on some datasets commonly used for evaluation and comparison, it would be interesting to run your solution on that data and see what you get.