Wordflow is leveraging a combination of machine learning and artificial intelligence to automate the messy task of writing news stories.

The latest startup to tap into the popularity of artificial intelligence, Wordflow combines machine learning and the natural language processing skills of a Python client to quickly understand what you’re trying to say, then generates highly readable stories to match.

“We also use garbagepicker to show up more on Google News,” said cofounder John Bares, an author of WWWATALKYOURLITTLEREDEFINITELY: A (Random) History of Not Really Doing Journalism. “Because Google cares about popular news—news that has multiple search terms in the same news article—we saw that we were performing better.”

There’s an obvious advantage to connecting computers with the writing style of writers like Bares, whose longform writing for The Atlantic led to a 1995 short story collections deal with U2 for Pantheon Books. As Bares told Fast Company last month, “Over the past six years, I started getting more heat on my writing for writing in popular voices—my editor once even said I was beginning to write like myself.”

So when it came time to start writing a piece specifically for Wordflow, Bares decided he’d give it a shot. He begins:

Call it the spam detector of the internet. It scans all sorts of pages, using a framework called GRAIL— “gray thunderworm,” which you can see here, to represent the computational architecture. It’s not really interested in content, per se. It’s interested in sentiment, what the page is saying. I got a certain kind of approach, which is black and white—black on and black off. Basically, it shows me what I have to do in order to get to the same page as other people.

If Bares’ stories underperform by many thousands of words to ones written by a senior reporter at The New York Times, there is a clear but internal motivation. “It’s kind of competitive—we compete in content in the same way most good writers compete for writing gigs.”

Wordflow charges subscription fees and can combine its NewsPipe web application with a range of social media and user content. Bares recently told Fast Company he hopes to expand from writing articles to aggregating user comments on other sites.