Why Kik Thinks Chatbots Will Kill Webpages

Kik’s new platform, which lets anyone create a chatbot, reveals some nuances about how messaging is evolving into a new UI paradigm.


2016 is shaping up to be the year of the chatbot: last week Microsoft announced a new platform for making chatbots; Facebook is expected to do the same next week; and today, Kik, the popular messaging platform that the company says is used by about 40% of American teens, is announcing its own bot-making platform.


Kik’s is a bit different from its peers, and the differences reveal some fascinating nuances about how messaging is evolving into a new user-interface paradigm to rival apps. For people who want to create a bot, Kik offers a simple API that allows them to specify how the bot will behave: What it can do, what it will say, and what responses you can type in reply. For users, the bots can be summoned into any conversation with an “@” mention. So you might be making plans with a friend depending on the weather, and bring forth @weatherchannel to give you an update on the forecast. Then, the bot simply leaves your conversation–a telling, human-centered detail meant to teach users that bots are here to help rather than to listen passively to their conversations. “It’s like asking a smart friend for an answer really quickly and not having them stick around,” explains Mike Roberts, Kik’s head of messaging.

In all, there are 16 bots launching today on Kik, including Vine, Funny or Die, Sephora, and the Weather Channel. The question, of course, is why this is happening at all.

Chatbots Solve Problems With The App Ecosystem

There are two trends leading toward chatbots: On the one hand, messaging is turning out to be the killer app of the smartphone. In 2015, according to eMarketer, 1.4 billion people worldwide used a chat app. On the other hand, the app model of smartphone development has been falling apart for years. No matter how many apps people have, they only use a few of them; Nielsen found that more than 70% of all smartphone traffic comes from just 200 apps. The promise of chatbots lies in meeting users where they already are–that is, chatting with friends–while at the same time offering a very low friction way to do new things. It’s a lot easier to “@” a chatbot than to learn about an app, download that app, wait for it to load, and then login with your information, and then try to remember what it was you wanted the app for in the first place.

Which brings us to what might be the smartest detail in Kik’s announcement: How users will find out about bots in the first place. One of the metaphors is familiar, and borrowed from the app store: Users can find new bots to fool around with in a “bot shop” that will soon list scores of bots, sorted by popularity and function. But bots can also be brought into any chat seamlessly with an “@” mention–which means users will teach other users what chatbots are there.

Unlike most apps, which have had a hard time spreading because of how hard it is to get another person to download the app, chatbot should be able to spread virally, through natural conversations. “It’s your friends who are going to teach you how to interact with bots,” Roberts says. “And you’re going to say, ‘What is that magic thing you just did.’ And then you have this whole new way to explore.” To Roberts’s thinking, the inherently social sphere that bots will inhabit is what will keep them spreading and keep users engaged, whereas apps tend to either boom or bust from the first day, and quickly see users disappear over time. And he expects teens in particular to cotton to chatbots, because of how deeply expert they are on mobile platforms. “In the mobile-first generation, they’re all power users. They use all the secret features, like how to plow through 60 snaps in a second,” says Roberts. “They pick things up, and they go deeper. They understand how to use them to make their conversations better or funnier.” Teens are also really good at showing 100 of their friends how to do something.


Machine Learning (Or Lack Thereof)

What’s notably missing from both the “bot shop” as Kik calls it is any sort of machine learning. To Kik’s way of thinking, bots aren’t meant to be conversational partners. They’re meant to do one or two really useful things—which means its enough to have developers program pre-canned queries and replies via the bot shop. There are only so many ways you’ll respond after the Weather Channel bot tells you what temperature it is; Kik thinks there’s more delight and engagement to be had if humans actually figure out funny or useful pre-set responses.

That might be the case, or it might not–it’s hard to say without using these bots more. But the greater benefit may be strategic. In avoiding the question of machine-learning for now, Kik is building something lightweight that it can easily refine, and assuming that AI will become a commercial platform they can eventually tap into. Why build in machine-learning, when it’ll be something you can get from Microsoft or Google?

When Microsoft announced new bot-making APIs at its recent Build conference for developers, it also announced that its machine learning technology would be available as little bundles of intelligence for developers to access—so developers might build their own bots, but lean on Microsoft’s image recognition to power new kinds of smart responses. Google and Facebook seem likely to follow suit, which means that those giant technology companies will likely lead the way in creating the intelligence that powers chat interfaces, while players like Kik will focus on developing the front-end experiences that will keep users happy—thus, there’s the potential of a frenemy ecosystem where players like Kik both compete with Google or Facebook or Microsoft, but also use its services to power their own apps.

For Google and the rest, machine-learning thus becomes a platform that keeps close those potentially disruptive apps that could lure users away; smaller companies potentially get the benefit of machine-learning they could never feasibly build themselves. Given the massive costs of developing any kind of machine learning platform–newly minted PhDs now command million-dollar pay packages–its likely that the tech giants have massive economies of scale.

Kik, for its part, wants to define an entirely new kind of affordance for the Internet. In Kik’s rendition of the future, bots will be, in a literal sense, all around you. Sit down in the stands at a baseball game, and there will be a sticker with a Kik logo on the seat back in front of you. Snap a pic, and a bot will appear ready to order you a hot dog and a beer, with the knowledge of exactly what seat you’re in. And so on, with any application you can imagine, from hailing a cab to your street corner or your groceries for next week. “We’re seeing chat become the new browser,” says Roberts, echoing the vision laid out by the CEO of Kik, Ted Livingston. “Bots are the new webpages.”


Which means it’s only a matter of time until a chatbot is listed on the NASDAQ.

About the author

Cliff was director of product innovation at Fast Company, founding editor of Co.Design, and former design editor at both Fast Company and Wired.