Conversational interfaces (texting, WhatsApp, even Twitter DMs) are where news readers spend much of their day, but newsrooms have been slow to adapt news delivery and engagement for those spaces, primarily because it's hard to scale a two-way conversations. New pilots and ongoing projects show that there's great potential for engaging audiences in messaging environments where they feel most comfortable—leading to a rich feed of stories and questions to drive reporting, but also as a "funnel" into a relationship with news organizations that lead directly to subscriptions and memberships. The bigger opportunity—and the foremost challenge—is transforming one-way, broadcast and publishing paradigms to two-way, more conversational ways of delivering the news and engaging with audiences. It's easier said than done: workflows need to be adjusted, stories need to be chunked out into components served up message by message, and—crucially—the organization needs to listen to what people are messaging back, and be ready to respond one-on-one and en masse. The obstacles are numerous, but the rewards—a more engaged audience, an efficient means for converting people from bystanders to supporters, a steady feed of insight and perspective—are worth the effort. And the advent of machine learning + natural language processing will quickly make supporting two-way conversations a great deal more efficient and intuitive.