NOTE This doesn't compile yet.

This still isn't optimal because in our code we use lists, so this
will fetch everything (using conduit), coerce into a list and then
process.

This fixes an issue when article title contained XML (like <b> etc).