We’ve recently argued that the Galaxy S7 edge may represent the culmination of the “Smartphone 1.0 era”. It won’t please everyone about everything — and no such thing will likely ever exist — but it’s indubitable how all of the cornerstones of a modern smartphone have been tackled cleverly by the Korean giant, and all its positives can even justify its hefty price tag. Actually delivering something that steps up the game in a significant way, in fact, looks like a very tough challenge. Perhaps we will have to wait until true, Project Ara-like modular smartphones show up before radically rethinking the way we look at hardware, but the software roadmap seems to be getting clearer, with a future studded with bots.
Microsoft is betting big on them as a major part of the future of computing, and so is Facebook. There is a case that could be made for them to become the new apps — and this certainly is how these two firms are pitching the concept. Chances are that Google will follow sooner than later, and I think that if the general idea of bots we have been so far given remains valid, the owning of a platform as popular as Android may leave the Mountain View behemoth with quite an interesting card up its sleeve, which could give them a notable lead in the upcoming war for bots dominance, were they to play it cleverly…
“Bots are the new apps” is essentially the message that was seeping through Microsoft’s Build conference two weeks ago. On Monday, Facebook kicked off its F8 keynote announcing the company’s full commitment towards bringing an army of bots into Messenger. Google’s big I/O fair is a little over a month away, and at this point I would expect them to put bots front and center as well. But first of all: what exactly are we dealing with?
Bots are essentially AI-fueled software applications that run automated tasks within a given framework, which at this point of time seems to be fixed almost entirely on messaging apps and “conversations”. Microsoft is betting big on Skype, while Facebook is naturally relying on the billion users actively using Messenger as their primary platform, with a few first examples (like the already-infamous “Hi Poncho”) mixing natural conversation and sprinkles of more classic GUI.
The premises (and the promises) are actually pretty big: imagine a single app capable of dealing with multiple requests without ever requiring you to leave it and hop in and out a myriad of other programmes. Wanna check the weather, order a double bacon cheeseburger, keep an eye on that Amazon package and send that PowerPoint over to your colleague? Just text all of that, and a smart bot on the other end will do all the dirty work for you.
It would be a dream for most brands and services, which could finally stop trying to figure out a way to ask you to install their crappy app. And it could be a dream for end users, too, which could finally free themselves of admittedly clunky processes and let smart virtual assistants manage their digital life. Well, as we said, that’s the future — hopefully. As of now, things are still pretty barebones, and not just in terms of availability, but actual convenience too.
Is it really better to ask a bot what the weather is when a quickly glanceable widget may result much faster? What if I’m set to buy a pair of shoes and, like in Facebook’s demo at F8, I am only presented with five options, which are likely not going to be able to narrow down the model I’m looking for? What happens if bots don’t give users the proper control they need and, like in the case of a news service, keep spamming you with texts of unwanted content? These are all challenges that conversational UIs and bots in general need to tackle.
Most of all, however, the big issue bot-pushing companies need to face is to either make these bots smart enough to truly understand human language — which seems a bit too far-fetched a challenge for now — or let the users understand the domain, i.e. the kind of language they need to use so that the bot can in turn comprehend. If a conversation keeps bouncing off because of the bot’s inability to understand what a user is writing, people will be naturally driven away.
So, to recap: apps are ubiquitous, and if bots want to sit alongside (let alone replace) them, they need to be just as omnipresent; they also need to prove themselves as an actually superior alternative to standard apps — and this seems to be the best case for their existence in a number of instances — and most of all be smart enough to overcome the challenges posed by human language: again, as clunky as apps may be, there’s little to argue with buttons and lists, where every action has an immediately obvious and unanimous reaction; it’s why “there’s an app for just about anything”, after all.
But where does all that leave Google? We know for a fact that the company is heavily invested into fields such as AI, machine and deep learning, but we haven’t quite seen them link those projects to, say, Android or specific applications. The clearest example may perhaps be found in the actual Google app, which has been increasingly putting down its roots into the OS, and I think that a quick, smart move may give them a significant competitive advantage, were bots to actually take off in the next few years.
Think about it: Microsoft and Facebook — the two tech giants that have so far showed significant interest into the concept of bots, showcasing their work at their big developer conferences — are companies who most certainly have the biggest interest in these bots’ mass expansion, because of their failure in the mobile OS space: if their strategy is distinctively ‘platform-agnostic’, it’s because it has to, as their only way of reaching end users is essentially reduced to — ironically — their own “old-school” apps, and thus all the limitations that come with installing them, as opposed to using something already embedded into the underlying system.
Sure, Facebook may have about a billion people using its service, but that doesn’t necessarily mean that all of them think of Messenger as a platform. In fact, especially in the West, messaging is not a prominent mean through which we interact with the world — much less than, say, Asia, where even apps like WeChat have already faced problems of this nature. A future made of bots is hence strongly subject to a present where the messaging-as-a-platform strategy is fully functional, and there doesn’t seem to be a strong indication of that just yet. Google, however, is lucky enough to be in an entirely different position.
A whole lot of people, in fact, already rely on a different kind of canvas, which is not strictly ‘conversational’ in terms of traditional messaging, but that works in a way that is very similar to the one so far fobbed off with bots. Google Now, the firm’s smart system that proactively provides the user with timely information thanks to pre-provided data, already allows you to insert a text- or voice-based query and have response. Sometimes it’s contextual and smart, other times it just returns a web search. But what if that became the canvas to interact with bots, rather than an app?
The search box sits atop almost every Android phone, and I believe this gives Google a clear potential head start over the other competitors. Again: albeit extremely popular, Facebook would likely still require you to open the Messenger app and then look for the appropriate bot, which doesn’t seem very quick nor intuitive, while Microsoft altogether lacks a mass-market smartphone app that has high chances of convincing people to switch over; why download Skype or Cortana when something else is already available, baked right into the system?
In a way, I would expect Apple to take the same road soon, but even if Siri were to one day mirror all of these capabilities on iOS, the number of devices involved would still be significantly lower than that within Google’s reach; and actually, given that Google Now lives well inside iOS as an app, some users may be persuaded to use it; given there are still no Siri APIs, time could be a deciding factor, and Google apps are indeed popular on iPhones and iPads. However, this could highly depend on the limitations the app would have inside Apple’s system, as well as all of the aforementioned problems, of course.
Nevertheless, Google should exploit this massive OS-based ridership — we’re talking ~1.5 billion devices on Android alone, not something to underestimate — and make it clear that such an implementation is available, easy to use and ready at one’s fingertips. If someone knew they could just type (or say) “call me an Uber” or “buy me a McWhopper” (please don’t) and have a quick, immediate response, I could definitely see people starting to use bots as an alternative to apps; especially single purpose ones. After all, at its core, this would be a very natural extension of what Now already does, but one capable of tapping into third party services’ bots other than the web.
Imagine using Now on Tap on a webpage where you see a certain product, and having Amazon’s bot ready in case you wanted to purchase it without ever leaving the page. Or perhaps think of a situation where a friend sends you their location, and regardless of the messaging app they are using, the system-embedded Now is capable of linking you with, say, Lyft‘s bot on the fly. Picture yourself shouting something like “show me the way to work with Citymapper” or “check that flight to Paris with Skyscanner” and having the system — with Now connecting with the appropriate bot each time — do all the work, perhaps even synced across all your systems, so to have results tailored to you regardless of the device you are using.
Lilian Rincon of Microsoft showed something very similar at Build, using Skype as a canvas; Cortana knew where she had to go and when, and thus prompted her with a hotel recommendation, which then led to linking her with the hotel’s bot and let her choose and confirm a room; these are all things that Google could manage just as well, retrieving data from our calendars and flights (as it already does), but being able to manage this ‘conversation’ inside of Now rather than Skype may have an entirely different outcome for the evolution of bots, hopefully leading us to that promised future sooner than we expect.
I believe that while apps are here to stay, bots may in fact represent the one change our smart devices need to become truly smart, allowing us to do not just an increasingly bigger number of things, but do them effortlessly and intuitively. Is this what Google is envisioning for the future of bots, Android and their integration? I don’t know. But I certainly hope it’s not too far from that.