Have you gotten the message?
Ben Brown has. The Internet pioneer, an old friend of mine, has long experimented with new interfaces. A decade and a half before millions of Pinterest users relied on “Pin It” buttons in their Web browsers to save images, Brown was inventing the bookmarklet technology they’d come to rely on.
Now the CEO of XOXCO, a product-design firm, Brown has become obsessed with a new interface: the message.
Specifically, messages in Slack, the popular team chat software that’s taken over workplaces from Manhattan’s media towers to the startup lofts of SoMa.
Brown’s company has just released Howdy, a bot for Slack which automates functions like gathering status updates for team meetings. XOXCO has also raised a round of financing led by Bloomberg Beta, the financial-software giant’s venture arm, and is focusing its business exclusively on Howdy. That makes the company the first developer, to my knowledge, to raise money specifically to build an app on top of Slack.
Here’s what Howdy looks like within Slack:
Where Developers Get A Lot Of Slack
Will Slack become a prolific platform the way, say, Windows was in the 1990s or Facebook and the iPhone were in 2008, spawning a range of new companies? So far, the company’s making the right moves. Slack CEO Stewart Butterfield recently said he’d rather make other apps work better on Slack than expand its range of built-in functions. And the company recently hired April Underwood, a well-respected executive who worked at Twitter, to run and expand its developer platform.
To date, most of the apps built on Slack that I’ve heard about have been for internal consumption, like Blossom, an app which recommends stories for New York Times social media editors to share. Blossom is just one of a range of apps media operations have built to automate their newsrooms, from streaming election results to checking colleagues’ schedules.
Other apps treat Slack like an extension of their existing services—one interface among many. Asana, a task-management app, has built an integration that lets Slack users send simple commands to create or update tasks in Asana, without having to leave Slack.
Fog Creek Software has taken that further, building Chatterbug, a simplified version of its FogBugz software for tracking bugs. While Chatterbug is built on top of FogBugz, Slack users can use it entirely within Slack.
What’s key here is that users are interacting with software bots in much the same way they chat with colleagues, turning work into an ongoing conversation.
And as bots get smarter, the distinction between our human coworkers and our digital ones gets harder to spot. And, perhaps, less meaningful.
Slack’s Underwood is bullish on bots.
“The potential for bots is really broad,” she told me. “I expect that expense-reporting software will have a bot in Slack. HR processes will have a bot in Slack. It’s not just optimizing small tasks. It’s as effective as having a team member.”
The Bots Take Over
Slack is far from the only place where you may not be sure whether you’re chatting with bots or humans.
Facebook is testing a bot-powered digital-assistant service in Messenger called M. Tell M what you want—for instance, to buy a shirt as a gift for your spouse—and it attempts to handle it with artificial intelligence. If the algorithms fail, humans step in.
Clara Labs makes a service called Clara, whose favored medium is email. Copy “Clara Turing”—the name is a nod to Alan Turing, the father of artificial intelligence—on a message about scheduling a meeting, and the Clara bot will try to find a time on your calendar that also works for your correspondent. As with Facebook’s M, when algorithms fail, humans step in. For $49 a month, Clara saves the tedious back-and-forth of checking calendars, and does it all over email—which is still how we humans end up booking most of our meetings.
WeChat, the chat service owned by Chinese Internet giant Tencent, has seen an explosion of bots on the service, including useful ones that help you hunt for jobs or buy clothes. But WeChat recently cracked down on some chatbots, including some developed by Microsoft, which has unsettled developers.
Clara recently booked a meeting for me with Matt Galligan, the former CEO of Circa News who’s now working as a product consultant. He sees response time as the big challenge for bot builders. In Slack, people will expect speedy replies, which limits the range of what bots can handle—especially if they need to fall back on human help, like M and Clara sometimes do. With email and messages, users may be more tolerant of delays.
Developers will have to figure out the nuances of each messaging platform. Facebook Messenger and WeChat have their differences; Slack is different from other team messaging services.
For Brown, though, Slack is a particularly bot-friendly environment. It doesn’t hurt that Slack’s built-in Slackbot welcomes users to the service, filling out their user profile by asking them questions rather than making them tab between text boxes on a screen.
“For people who get Slack, [Howdy] is going to feel like ‘more Slack,'” Brown said. “This is just an extra part of the experience.”
Like a free-flowing conversation, Slack is an “unstructured environment,” according to Brown. That’s fertile soil for developers to grow in.
And there are implications beyond workplace productivity. Consider the vast amounts of data generated by connected devices, at home, in offices, and around cities. What will be their means of communicating data, insights, alerts, and other information? Increasingly, we’re drowning in apps and websites. We need new, humanistic interfaces—like a conversation. It’s just a matter of time before our bots are smart enough to get the message.