Product trends with Intercom's Paul Adams and Emmet Connolly

Things like chatbots, voice UI, conversational commerce, machine learning, moving from screens to systems – all were discussed and debated anywhere and everywhere, including our own blog. We at Intercom also had a handful of massive launches in 2016, including Smart Campaigns, a new Messenger and Educate, our knowledge base product, and that has us reflecting on a year’s worth of lessons learned.

To make sense of the past year and dig into where product and design are headed next, I hosted a roundtable discussion with Paul Adams, our VP of Product, and Emmet Connolly, our Director of Product Design.

If you like what you hear, check out more episodes of our podcast. You can subscribe on iTunes or grab the RSS feed.

What follows is a lightly edited transcript of the interview, but if you’re short on time, here are five key takeaways:

1. Given today’s technology, chatbots are best left to handling computation. Things that require empathy or emotion, on the other hand, are still better handled by a human.
2. From Airbnb’s launch of Trips to Instagram’s addition of Stories, products built as systems rather than a set of screens became more prevalent in the past year. As new uses are demanded for your product, your system will have to expand.
3. Breakthrough products target existing behaviors, rather than asking users to break from the norm. 2016 featured two prime examples of the former: Snapchat Spectacles and Tesla’s solar tiles.
4. Product teams must make a philosophical shift after they launch a product. As the team enters iteration, every previous decision is back on the table.
5. Defining conversational commerce as sending texts to a bot is simply too narrow. Product builders must expand that view and look at their product as an ecosystem with many endpoints – and messaging is just one of them.

Des Traynor: Today I’m lucky to be joined by Paul Adams, our VP of Product, and Emmet Connolly, our Director of Product Design. 2016 was marked in a lot of ways by bots. We had our own opinions, and we had our own experiments, as did the entire industry. Is the future of product design really gonna sit inside a chat bubble?

Paul Adams: Both Emmet and I wrote a lot of blog posts about bots over the year, and we built a lot of bots too. Some of them saw the light of day, some didn’t. We learned that bots are overhyped.

For very human things like empathy and emotion, bots are terrible.

What we didn’t realize is that bots do work for a very specific set of use cases that are probably narrower than people first imagined. There was a crazy AI vision of the future, where bots are as intelligent as humans, and our biggest realization was that bots are good at some things, and humans are good at other things. Bots are really good at computation. Bots are basically simple computers, so if you need to ask somebody what your next bill was gonna be, a bot can calculate that far faster than a human, who’d have to look up the system, find your account, look at the UI and find the number. For very human things like empathy, emotion and reading between the lines of what someone’s actually trying to say, bots are terrible at that, given today’s technology.

Des: Emmet, from a design perspective, it sounds like you’d have to spend half your time dealing with whether or not the bot knows the answer. In the majority of cases the bot’s probably not going do a good job, right?

Emmet Connolly: We have a system whereby a human or a bot could answer your question, and so it becomes more of a rooting problem than a problem of, “What do I do in this failure case where the bot doesn’t know the answer?” If the bot doesn’t know an answer, or can’t provide a great one, then the human should provide the answer.

Paul and Emmet’s teams designed Educate, Intercom’s new knowledge base product, for bots to supplement a human customer support where appropriate.

A lot of the pitfalls we saw this year were use cases where people building these bots were over-promising what they could deliver. The technology for an English-language level conversation really isn’t there yet, and that has plunged us into this trough of disillusionment. That’s also a good place to be, because it means that we’re getting real about what’s actually possible. If 2016 was the year of hype around this, we could actually see a lot of real life, useful tools and products emerge in the next year.

Des: It’s really convenient, the way these things pick whole years in which they’re going to experience these iterations. We see bots that pretend to be humans, like, “Hi, I’m Barry the airline bot, how can I book you a flight.” And then you see bots that are blatantly bots, like, “I’m the little operator bot, and I’m going to point you in the right direction.” You said the idea of trying to humanize these bots isn’t something that we want to do at Intercom, but what’s the general thinking there?

The degree to which you personify the bot evokes a very different reaction in the end user.

Emmet: Our thinking has actually evolved a lot as we’ve tried out a lot of the experiments that Paul mentioned. Initially the thing that seemed most obvious to me was, “Hey, these are friendly little robots that can interact in your conversation. Let’s make them be tiny Pixar characters.” That’s not what resonated with the users that we put our early bot iterations in front of. The nuance of tweaking a little bit of the language or the degree to which you personify the bot evokes a very different reaction in the end user. Some of our early experiments had people saying, “Hey, I’m ‘bot name’, I’m not a real person, but I have a character.” People didn’t like that at all, because hey felt slightly duped by it. They thought they were here to talk to a person. If you can insert a level of automation and, “Hey, I am an automated bot that’s here to speed up the process,” then people can see the value in that, and it doesn’t feel like a bit of a bait and switch.

Des: We had a command line once upon a time, “Write in the exact word and you’ll get the exact answer.” Do you think people’s behavior changes when they know they’re talking to a bot? Do they still continue the formalities and the civility and the, “Hey, I’m curious about…”, or is it just like, “Flights please”?

Paul: For me, this thing is a scale. At one end of the scale is a command line interface, where it’s clear that you’re talking to a computer. People don’t actually, in many cases, turn around and ask themselves, “What is a bot?” There’s no actual common definition. We, in one of our blog posts, said that a bot is a simple computer program that executes, and then to Emmet’s point, you can give it a face, you can give it a name, and you can make it more or less human-like.

If the command line interface is one end of the scale, at the opposite end is what Facebook was trying to do, whereby you didn’t know if you’re talking to a person or a bot, and clearly there somewhere is the uncanny valley. We didn’t get close to the uncanny valley. We were far down the path of this is clearly a computer program. The minute it started pretending it was anything other than a computer, people reacted very negatively to it.

Bots are most effective today when users know exactly what they’re interacting with.

Des: So it comes down to honesty and transparency?

Emmet: Partially. I also think part of the 2016 exuberance from bots was around this sense of, “Check out this very simple use case that I made up, screen-shot, and put in my blog post,” and people with a sense for product said, “Wow, that does seem like a real new simple way of doing things.” Then when you try and actually build these things, you realize that typing human, English-language sentences to a robot isn’t how you want to interact with a robot.

If you give people a blank input field, it leaves a lot of room for the person to be rude to the robot or type pseudo-commands or friendly questions to them. There’s an evolution of the input on the end user’s side that isn’t quite so English-language based. Maybe some of that standardizes around how you might be given a set of pre-canned responses that you can send to the bot, which is faster than typing a whole sentence. It’s easier because you actually know what you can say to the bot, which is another problem these blank input fields provide. If I’m right about that, it’s possible that that question of, “What should your tone be when interacting with a bot?” would go away.

Des: If you follow that thread all the way, it’s hard to see how that doesn’t start to look like buttons that you click. It starts to look like UI.

Emmet: The most salient characteristic of conversational UI is not necessarily that you’re typing raw text in at the bottom of the screen, it’s that you have this back and forth log of commands. It goes back and forth, and there’s more ways of sending a command to the computer than typing it into the input field at the bottom.

Des: And maybe there’s not a predetermined order the command should be received in, and they would be in a form or something similar.

Paul: Right, and to play at both sides there, there’s also more ways than having this nice grid layout on a screen, with a left-hand nav and a button at the bottom, and all the other common ways in which we’ve come to expect products to work.

The rise of systems

Des: Paul, you spoke a lot this year about the idea that products should be thought of as systems rather than just a set of screens. How has that played out a lot in 2016? Have you seen a lot of new systems emerge?

Paul: Most people building software are actually building and designing systems already – they just may not necessarily realize it. Systems are these broad networks of things that are related and connected. Uber is a simple example. Uber is certainly not a software company and certainly not an app company. Uber is this ecosystem of drivers, passengers, inventory, all sorts of things. Over time as that system emerged and evolved, Uber added things like surge pricing. Other parts of the system like price and availability and even where people drive – the whole thing changes. You change one piece of the system, and other the pieces of the system change.

One example from this year that I use is Nike+. Their app was totally redesigned this year. It used to be called Nike+ Running. The redesign is now called Nike+ Run Club. That’s a sort of branding distinction. It’s like, “Oh, it’s a club. There’s other people. Oh, it’s a system. All right. There are connected things happening in here.” If you go into the Nike+ Run Club app now, you can go on runs, track your mileage, all the normal stuff that running apps have. But they’ve paid a lot more attention to the system of runners – runs with other people, meetups in different cities, your running shoes. Nike+ had a lot of these components but they’ve certainly doubled down on an idea that this is actually a broad ecosystem far beyond the app. The app is a conduit to other things happening in real life.

Des: What about from the design perspective, Emmet? Have you seen any new things emerge in this regard?

Emmet: Airbnb went through a similar transformation this year. The objects or actors in that system used to be things like guests and hosts and housing accommodation. Now they’ve added Trips, very much broadening their scope. One would assume they’ve had to evolve their underlying system to cater for these new things that they’ve added. That suggests to me that the system will tend to expand as new uses are demanded.

If you looked at the Instagram app earlier this year, the big button in the bottom middle would open your camera and you would post that to your stream. Now they’ve added Stories and almost flipped what they’re about. Now the button at the bottom is to upload a photo you’ve taken before. That’s your curated feed. Honestly, it’s probably what Instagram was always about.

Snapchat is the possibly unfair, hard to avoid comparison when you look at how Instagram has evolved their system. Snapchat in itself is an interesting example, because they came from a very chaotic place earlier in the year and they seem to be rationalizing a lot of concepts. You could say that Instagram and Snapchat are meeting in the middle having come from very opposite places.

Targeting existing behaviors

Des: Instagram started with a very clean system. It was very obvious you had your own stream and other people’s streams, and that was it. Snapchat started from, frankly, who knows where. It was the most aggressive disrespect of systemic standards of software. I heard it argued a while ago that Snapchat’s an unconventional UI was actually part of it’s genius – it made the product somewhat viral because everyone had to teach each other how to use Snapchat.

Snapchat themselves have had a interesting year. Aside from blowing up, they also released hardware for the first time. They took a really different approach from the likes of Apple, who are trying to get people to either replace or start wearing a watch, or Google Glass, who are trying to get people who don’t wear glasses to wear glasses. Google actually had nothing to offer people who did wear glasses, which left them in a difficult situation. Snapchat said, “Screw all that. Let’s just sell some sunglasses.” To me, that approach lines up, because they’re targeting an existing behavior, sunglasses, with a better product. That is usually how good products happen. How do you see it?

Paul: My take is pretty simple. These products are attempts at major breakthroughs for how people live and act, and the ones that are successful don’t necessarily try to change people’s behavior. They understand how people behave, act and think, and the new thing that emerges is congruent with that.

This idea of building bridges to the future was very prevalent when I was at Google. Things like Google Glass and Google Wave, these were products that didn’t really build bridges to the future. They were so different and forced people to act in totally new ways. People just didn’t know what to do with them. Whereas Snapchat’s glasses are sunglasses and that’s it. They’re sunglasses with a camera, and people wear sunglasses.

If you actually watch the launch video for Snapchat’s glasses (above), it’s not people flying around on skateboards. It’s literally a shot of normal life except they’ve different sunglasses on. Google Glass’s video had all this augmented reality. It was like a science fiction movie. It’s obvious to me that the winners are the ones who don’t try to radically alter society from the get-go.

Des: The bridge is the right concept for that. Snapchat could roll out a next version with a screen inside the lens that lets you see filters applied, but obviously they have to get there.

Paul: Look at the development of the horse to the car. The very first things that emerged were basically horseless carriages. You can actually see a step by step progression.

The winners are the ones who don’t try to radically alter society from the get-go.

Fast-forward to today and look at Tesla. The Tesla car doesn’t need to look like a car. It doesn’t have to have an engine at the front. The chassis is the mechanics of the car, but it still has something that looks like an engine and something that looks like a trunk. That’s because people need this bridge. They’re not going to buy a Tesla if it looks radically different to a car.

Emmet: I wonder if this bridge to the future concept is especially true for consumer products, where no one wants to seem like a weirdo on a pseudo-futuristic thing. The car is an interesting example. If you think about a thing that’s likely to somehow achieve mass adoption at some stage over the next ‘x’ years, it’s self-driving cars. There will be a gradual acceptance of the general public to self-driving cars and your willingness to take your hands literally off the wheel and trust the computer, because that’s a consumer use case.

If you think about self-driving trucks, fundamentally the same product, I would anticipate there will be an almost overnight switch as soon as it becomes possible from a regulation point of view, mostly from a bottom-line finances. Some CFO of a massive company is going to make the switch overnight to self-driving trucks, and likely the entire industry will follow within a very short amount of time.

Shifting from product launch to iteration

Des: You both spent a lot of the year working on our new Educate product, which is Intercom’s take on a knowledge base. The product’s launched, and we’re in a different mode. What happens after launch for a team?

Paul: It’s been fascinating because the mode the team needs to work on almost the minute after the product launches is different to the one that they had before launch. This happens overnight. In the lead up to launch, design decisions are antagonized over for a long period of time, but they’re at some point locked down. It was like, “Do not open this conversation again. Do not increase the scope of this thing. Narrow, narrow, narrow, narrow. Get down to the hit list. Knock it off. Knock it off. Knock it off. No opening up of old things.”

The minute we launched, suddenly, that all goes away and everything’s on the table again. It’s the nature of all software that when you launch something, you’ve got loads of stuff wrong. Loads. Even if you’ve run a great beta. Suddenly, every single decision you made is up for debate again. It’s very hard to suddenly accept that we might change it all and do it all a different way and iterate.

A sign of a great product team is not what they ship in their version 1.0. It’s version 1.1.

Emmet: There’s almost a tendency to become more conservative. You’re thinking, “We got it working. We finally got past the finish line,” and now some people are using it and liking it. You have to be extremely brave at that point and say, “There’s a huge amount of input and information that we didn’t have before.” A sign of a great product team is not necessarily what they ship in their version 1.0. It’s that version 1.1. That’s when you really see how great a product team operates and thinks, how agile can they be in adapting to the new world where the rubber hits the road and their product is out there and they’re getting feedback.

Paul: Instagram impressed me in its early years because of it’s simplicity. I was at Facebook when Instagram was acquired and the Instagram team was incredibly disciplined. There was pressure from all over the industry to add this or that, especially the minute Facebook acquired Instagram. “Why doesn’t Instagram have this feature if Facebook does and Facebook knows it works. They should just add it.” The team was incredibly disciplined to keep Instagram really simple.

A lot of Instagram’s success in the early years was because it was so accessible and simple to understand, and they stuck to this core use case. Credit to them as Snapchat has moved the industry on. They’ve been able to adapt and that is the sign of a great team, a team that can change.

Des: They both represent proof that you do not need a large product footprint to have insane engagement. They’re both relatively small product footprints. They have managed to get stronger, better, more engaging and more addictive without actually adding 15 more screens or 25 more workflows, which is stock B2B SaaS philosophy.

What’s to come in 2017

Des: We’re going to do a lightning round, so short answers here. Voice as a user interface, will it be relevant in 2017?

Paul: Absolutely. Huge.

Emmet: No doubt.

Des: Relevant in B2B as well?

Emmet: Too early to say.

Des: Conversational commerce – overhyped, absolute bullshit, very meaningful?

Paul: Not overhyped, definitely a big deal, and probably has a bad name.

Des: Conversational commerce is way too broad.

Emmet: It still holds great promise for very specific use cases, but if you expect that all of commerce will move into your messenger you’re probably sorely mistaken.

Des: Virtual reality – will it be relevant in B2B and B2C in 2017

Emmet: Too early. Its’ time has come for gaming and that’ll be exciting to see what the possibilities are. It doesn’t feel like all the pieces are there yet.

Des: Worst trend in product design today?

Emmet: The silver bullet mentality, whether it’s conversational commerce today, bots throughout this year or shiny buttons fifteen years ago. There’s always something and you’ve got to consider whether it’s a trend or an actual new building block.

Paul: AI. No one actually can define agreeably what AI even is. When most people say AI they mean if, then, this, that statements. Just take some humble pie and agree that we’re not there yet.

Des: What was your favorite new product of 2016 and why?

Emmet: I have a very un-sexy answer for this – the Belkin WeMo light switch, which has removed this very small but noticeable friction from my every night of shutting off all of the lamps around the room. At Google there used to be this concept of, does a product pass the toothbrush test? Would you use it at least a couple of times a day? This removes a very tiny thing but if you can make a tiny improvement a couple of times a day in every day of your life, that’s decent. I will say that from a product point of view it’s still a total mess. There’s actually no system behind how home automation works to really tie it together so you have to Sellotape all the bits together.

Paul: One of the most ingenious, amazing inventions of the past year was Tesla’s roof tiles. Don’t put these big panels on your house, just make tiles that do that instead. It harks back to the idea of building a bridge to the future. [Would you rather] put this foreign looking sci-fi object on your roof, or just swap out your tiles?

Tesla’s roof tiles give users a solar option without requiring a radical change in behavior.

Emmet: If you want to consider a system of interacting elements, the ultimate example must be Elon Musk’s zoomed-out plan where one of the things he needs to put in place to get humanity to Mars is better roof tiles. Like who would have thought, right? But all of these interlocking strategies of these companies fit together to meet his ultimate goal.

Des: I’m going to bring it back way down to Earth and say my favorite product of 2016 has been Bear, the note-taking app. Biggest lesson you’ve learned about product building in the past year?

Paul: We’re a company that has valued getting value out to customers as fast as possible, scoping small. But we did three big projects this year. Each, at the time, was actually the biggest projects in our history. Our Smart Campaigns launch in May, our new Messenger in August, and then Educate, in December. I learned that you shouldn’t do three big projects all at the same time because it’s just really hard.

Emmet: I would agree, specific to our own experience doing those three big projects in the year. But when you’re taking on that amount of work you have to figure out where and when to sweat the details. If you’re going to obsess over every tiny thing then you’re going to get almost nowhere. It’s another one of those art versus science things – there’s no hard and fast rules but you get a better sense over time. We taught ourselves that lesson of when to make that trade off many times this year. When to let go and live with it for now versus when to make sure that you polish the hell out of one specific thing.

Des: Lastly, what do you think will be most relevant in product design and product creation in the year ahead?

Paul: As much as we’ve criticized bots in this chat, I do think that in 2017, we will see really, really useful applications of … I’ll say bots, but really it’s automation. Simple automation via conversational interfaces is one way to describe it. I think we’ll see real applications of these things, situations where computers are better than humans at doing things people need them to do and I think we’ll see a really quick maturation of the use cases for these bots shine.

Emmet: Defining conversational commerce as sending texts to a bot is a bit narrow. Broaden it slightly and think “I’ve got a service, and there are many, many endpoints, whether it’s an Amazon Echo, a voice UI, or Facebook Messenger.” It’s system-level thinking. When is the appropriate place to surface my functionality to someone? Do they want to speak, or type, or do something else that’s most efficient or comfortable for them? Think about your product as more of an ecosystem with many endpoints; it could surface in many places and not just in messenger.

Paul: The interesting thing is it reduces the idea of UI’s, wireframes, and screens to being like a window into a system. That window will change over time, as will watch interfaces, voice interfaces and bot interfaces. Product folks should be thinking about systems. Not just because it’s cleaner, but because it’s the best way to have sustainability.

Emmet: Things like the taxonomy of the site and thinking about this as a structured entity are really important. As an industry, a lot of the attention in the past ten years or so as we became very mobile-centric went to one of screens, and a small number of screens, and unbundling functionality into very small apps. I think we’re moving a bit back. Not back to where we came from, but a lot of those skills that served us well when we thought about things like information architecture and products as abstract things will serve people well in this new world.

Paul: X.AI is a really interesting startup. Their main interface is email. Then there’s Magic, and their main interface is SMS. We were in a period where the mobile screen was the dominant pattern that everything else emulated from, and we’re fast moving beyond that. Echo and Alexa, these things are taking off. Our world is going to look quite different this time next year.

Des: That’s technology for you. Thanks very much folks and thanks everyone for listening.