When do bots beat apps? When context and convenience matter most

by Peter Rojas


There's a question I've been asking founders when they pitch me their bots: "Why is this a bot and not an app or a website?" I've been fascinated by the explosion of developer interest and startup activity around chatbots and am pretty bullish on their potential as both a user and an investor, but it's important to take a step back and think about why we're all so focused on conversational interfaces. That means understanding that is often context that determines whether a bot is more useful than an app or website. 

It's undeniable that a huge part of what makes chatbots so attractive is that they promise to make computing more natural by stripping away graphical-user interfaces and instead let us simply ask for what we want. We're not there yet -- and won't be for a while -- but today's chatbots show us a glimmer of that future, one where we can simply speak or write something and the computer will understand what we want and give it to us. 

That gap between what chatbots can do now and what we know they'll be able to do in a few years can be enormously frustrating -- anyone who has tripped up Alexa or Siri by not phrasing a request quite the right way can tell you that. Certainly one of the biggest hindrances for chatbot adoption will be the amount of syntax you need to memorize in order to interact with them. Eventually, advances in natural language processing and AI will get us past all that and to a place where you can ask a computer for almost anything and it will be able to know what you mean. 

Although we're not there yet, today's bots can still be useful despite not having anything resembling true AI. We're already starting to see how a whole host of everyday tasks can be accomplished by chatting with an automated agent. Even if that bot isn't especially intelligent and requires you to employ a bit of syntax in order to interact with it, there's something about being able to have a quick conversation with a chatbot to get something done which can seem sort of magical.

The ease of conversational interfaces might be the primary driver of adoption, but there's something which has been missing from the debate over chatbots vs apps, and that is that the context in which you're trying to do something matters a lot, maybe even more than ease of use. A large part of what makes chatbots so compelling is that conversational interfaces have the unique ability to be integrated into the context we're already in. That's why chatbots make the most sense when the cost to switching contexts is high. Whether it's information, content, or a service, being able to ask for what you want via chat isn't merely about making something easier to do via an app. It's also about making it easier to have that interaction while having the minimal amount of disruption or interruption to whatever it is you're already doing. Because of this, a chatbot doesn't necessarily have to be easier to use than the corresponding app. It just needs to be more convenient to use given the context in which you want to accomplish a given task.

Let me explain what I mean. I have an Amazon Echo in my kitchen. Probably the number one thing I ask for each morning is the weather. Could I just pull my phone out and check the weather there? Sure, but it's much easier when I'm making breakfast to just say "Alexa, what's the weather today?" I don't have to stop what I'm doing and switch modes. Similarly, Slack bots make it easy to do whatever you need to do inside of a collaborative work environment rather than doing it outside and then pulling that information (or whatever it is) back in. Integrating your service with Slack as a bot lowers barriers to adoption because it can be used right inside the conversational flow which the business is already in. 

Chatbots are a bet that we are going to be spending more and more of our time within messaging apps like WeChat, Facebook Messenger, Kik, Telegram, etc and that it will be easier to access the services we want via a bot within those apps than to jump into another app or use the web. This is already proving to be the case in China, where a huge number of WeChat's 650 million users spend an enormous amount of their time within the app. They're not just spending that time chatting with friends, they're also accessing a whole range of services within WeChat itself, including shopping, virtual friends, games, etc. 

This is why Facebook is making such a big push with their forthcoming bot platform for Messenger. What they're rolling out will presumably make it possible for pretty much anyone to offer their content or service within Messenger. But while a lot of that will be the same stuff you can get via an app or via the web, Facebook isn't just wagering that the ease of interacting via a conversational interface will drive uptake of chatbots amongst its 800 million users. Ultimately they're doing this because they believe that the convenience of chatbots will get people to live inside Messenger in the same way that WeChat users live inside that messaging app. It's their way of making an end-run around both iOS and Android as app platforms by bringing all those services within Messenger as chatbots -- and thus onto a platform which Facebook controls. Uptake may be a bit slow as first users get accustomed to interacting with chatbots, but it's not hard to imagine user behavior changing over time, particularly as richer, more app-like UI elements get incorporated into chat platforms that further blur the line between chatbots and apps.

If bots win, it won't simply be because of their ease of use, but by the sheer usefulness of being embedded within the context in which we are already working or conversing, thus saving us the hassle of having to jump into a separate app or website to accomplish what we want done. Ultimately what will get us chatting with them is the convenience of having them there when we need them. You shouldn't have to go out of your way to get what you need from a bot.