Chatbots: an interview with Chris Messina, inventor of the hashtag

Are chatbots going mainstream in 2017? As designers, what are some of the challenges of designing conversational experiences?

Following up our report on the State of UX in 2017, we are interviewing designers who are big thinkers for each one of these important themes for the design community in 2017. Let’s keep the conversation going!

Chris Messina is a product guy, inventor of the hashtag, ex-Uber, ex-Google, and friend to startups. Oh, he also created his own personal bot (say hi) and set some time to talk with us about how he sees the future of bots… and humanity.

Everything is a conversation

“Chatbot” is one of the hottest terms in our industry right now, and we are pretty confident you are going to be building one quite soon — if you haven’t already. But what does the future of Conversational Interfaces look like?

If you’re reading this article, there’s very little chance you haven’t heard about Conversational Interfaces in 2016. At, we’ve written about the technical and social challenges of designing conversations, helped designers who wanted to get started in that space, talked about prototyping bot experiences, and even curated some of the best chatbot experiences we’ve seen this year.

Every interface is a conversation

Essentially, a Conversational Interface is any user interface that mimics chatting with a real human.

But stepping back for a moment: isn’t every interface a conversation between the user and the machine?

Think of the most common apps you use everyday. Like hailing a cab.

First, you tell Uber that you need a ride. Then, it asks you where you are, and once it has found a driver, it tells you the time estimate. When the ride is over, it asks you how it went. And you tell it your opinion by clicking on the stars and rating the ride.

Uber: a conversation about your latest ride

Traditional interfaces (the ones we design every day) are quite similar to a conversation — that just happens to manifest as buttons, menus and other interaction patterns. With Conversational UI the structure is the same. But instead of buttons, menus, and stars, you tell the machine what you want using words. And emojis ?

Conversations will only get louder

“Chatbot” is the next big buzzword in design — and our industry is seeing a lot of interest from companies in exploring that space. Automated, conversational experiences allow brands to inspire, communicate with and serve their customers right where they are, in a much more scalable way.

Order pizza from Facebook Messenger with Pizza Hut’s new chatbot

Apps like WeChat have become the central destination for a plethora of services in China. Over half a billion people use WeChat, and it touches everything — from consumers communicating with friends, to sharing their daily moments, to buying everything from food to paying credit card bills. It’s IM, ecommerce, banking, dating, gaming and marketing rolled into one platform, where you can shop, order food, book doctor appointments, find parking spots nearby, book hotels, hire maid/nanny/babysitter, hail a taxi, and so on. All through conversations — and mini-apps that run within those conversations.

Messenger, Kik, Slack, and many other messaging platforms have been working hard in 2016 to expand their capabilities and allow for similar experiences through conversations.

Not to mention voice interfaces — Siri, Alexa, Google Home, and so many others — a natural next step for chatbots, and a business opportunity that will inevitably affect the way you, as a designer, think about products and services in the near future.

The interactions of the future are not made of buttons.

Will 2017 be the year where companies shift some of their primary experiences to a chat format? Have we found real use cases for it, or are we just following the hype?

Chris Messina is a product guy, inventor of the hashtag, ex-Uber, ex-Google, and friend to startups.

Bots have been around for decades. Why is everyone talking about them now? What has changed?

Chris: It’s true that automated assistants — or the concept at least — have been part of computing lore from the beginning. In fact, the first utterance of the word “robot” came in 1920, thanks to Czech playwright, novelist and journalist Karel Čapek. Ever since humans have built machines, we’ve dreamed of animating them with intelligence and turning them into companions or helpers. There have been numerous attempts at this over time (remember Clippy? Or C-3PO?) but only now is the technological underpinnings and social behaviors aligning together to make it possible for bots to enter into the mainstream consciousness.

Technologically, we’ve reduced the costs of massive computing operations as the number of vast datasets has grown. With so-called “cognitive computing”, computers are able to simulate patterns of human thought and memory, and when applied to social computing applications, get pretty darn close to something that passes for being “humanlike” (think Alexa or Siri).

Furthermore, the number of people using computers in various forms continues to increase around the world, and especially for newcomers, the natural mediums for bots — voice and messaging — are far more accessible than their graphical counterparts, at least in broad strokes. For example, it’s much more likely that my 11-year-old daughter will be willing to ask Alexa to play Kidz Bop for her than trying to sign in to my Spotify account, search, and then choose an album when all she wants is music. This pattern is likely to repeat for a large number of people just getting started with their digital lives — and we tend to discount those users because their experience is so removed from our own.

Finally, besides the technology and the familiarity and accessibility of voice and messaging to most mobile device users, there’s also this crisis in engagement where conventional apps are just really hard to grow adoption and sustain usage.

Whether it’s app fatigue, device limitations (i.e. storage), or adverse networking conditions, most people never download any apps.

The combined friction is just too great, and yet businesses want to be in touch with their customers, and their customers want convenient, easy access to their favorite brands, businesses, and services.

Messaging in particular is extremely familiar and accessible, and so the next challenge for these brands, businesses, and services is to scale their presence in that channel.

That’s where bots (automated systems available via conversational contexts) are essential, and becoming increasingly important to future-leaning companies that want to be present wherever their customers might be.

As any new technology, it takes time for designers to figure out the most relevant use cases to apply it for. What are some of the use cases where bots are seeing the biggest success right now?

Chris: It’s not unusual to try to cram the previous generation’s content into the latest technological medium. It’s happened over and over throughout time, whether you’re talking about bringing the newspaper to radio, radio to cinema, or cinema to television. The shift from the web and apps to messaging or voice will follow a similar pattern until people start to experiment with, take risks on, and ultimately develop experiences that are inherent to and optimized for the messaging and voice contexts. Bots give us a slight advantage, because they’re not entirely new — but what is new is that now there are highly competitive and increasingly expressive platforms that offer new functionality and integration points that previous era bots had to compensate for with imprecision, wordiness, or arcane commands. As these platforms are still in their infancy, designers should really dig in and learn about each platforms’ UI primitives, capabilities, and perhaps more importantly, try out a bunch of the bots.

Designers should really dig in and learn about each platforms’ UI primitives, capabilities, and perhaps more importantly, try out a bunch of the bots.

It’s still super early and the relative ease of building a bot belies how hard it is to create one that’s actually any good. I’ve seen a couple that are pretty okay, but it’s rare that a bot will blow me over. We need more of those.

Is it just a hype? Or is 2017 the year bots will go mainstream?

Chris: When I declared that 2016 would be the year of conversational commerce a year ago, I was identifying an emergent trend that messaging was about to become a great new context for businesses to reach their customers and vice versa. The ensuing hype (along with the dozen or so messaging and bot platforms that were launched by all the major tech companies) seemed to suggest that we were on the verge of WALL·E level artificially intelligent agents that would be at our beck and call, all available over our favorite messaging app or listening computer. But, clearly, by the end of 2016, there was still a long way to go and much work to be done.

From my perspective, and reflecting on the fact that the iPhone is 10 years old this year, I think we’ve got at least 3–5 years before bots are “good enough” to really seep into our everyday experience without calling attention to their presence. In other words, it’s no longer remarkable for someone to pull out a touchscreen-based supercomputer from their pockets in casual contexts these days; to get to the same level of comfort and ease with bots, I think we’ll need a few more years for the use cases to evolve and the technological foundation to become more expressive, adaptive, and personal.

It’s no longer remarkable for someone to pull out a touchscreen-based supercomputer from their pockets in casual contexts these days; to get to the same level of comfort and ease with bots, I think we’ll need a few more years.

So no, bots won’t go mainstream in 2017; we’ll see a lot more of them for sure, but the real shift to watch for will be in how products and services become more conversational in nature and design, and increasingly work through whatever interface (visual, voice, or messaging) we happen to be using at the time. Those who go truly cross-platform (in the broadest sense) will have a huge lasting effect and competitive advantage.

In your opinion, what is the biggest design challenge of designing conversations? What are some things that designers should keep in mind?

Chris: I’ll admit it: bots and voice computing are fun to design and build for. You can iterate really quickly and it doesn’t take much effort to launch a bot. But, as it is with people, designing a bot that’s interesting takes real craft and skill. You’ve got to go much deeper than just style — substance and adaptability really matters! And while it’s true that many bots should really stay focused on one or two core competencies, the art of failing gracefully will be on display in a way that it’s just not in most screen-based interfaces (where you can just offer a link to go back home).

Writing conversational dialog for a bot is different than just writing a script because the cadence and timing and tiny details like typing indicators matter. It’s a much more intuitive design process that’s hard to outright automate or leave to writers or content marketers. The combination of synchronous and asynchronous modes of conversation require you to hold state for each user longer, and to remember things from previous sessions that you normally don’t need to in web or native apps. So, you really have to get into the real world of your users, and understand the flow of their day, and when your bot might be the most useful to them, and how chatty (or not chatty) you bot should be.

I know that I’m increasingly interested in improv and comedy acts and reading comic book dialogue to see how other pros have figured out how to make conversation engaging and to avoid dead ends. It’s really about keeping the conversation going, constantly evolving and tuning your approach to conversation design, and building up a real sense of rapport with your user over several sessions. It’s a fascinating and challenging context to design for, but one that I believe will be increasingly rewarding for those who really dig in and go deep.

Thank you Chris Messina for your participation. See you in the next interview.

Author: Fabricio Teixeira

Collect by: