Blog postThe Meaning of Bots


If you haven’t already noticed, bots are on the rise and we take a look at what they are and what they mean. Bots are new platforms for developers, they are an interface for users, they represents a whole range of innovative products and they are also a new channel for brands and enterprises to interact with consumers. This is a roundup of the current state of bots.   

Chatbots. You might want to digest that word for a little while. Although the concept of bots has been around since the 1960s in various forms, it did not really come to exist in the public consciousness until just recently. It was made (in)famous by Microsoft’s freewheeling racist twitterbot Tay, almost as unpredictable as HAL 9000 in Stanley Kubrick’s 2001 A Space Odyssey. A sexy female and friendly bot, Samantha, later turned up in Spike Jonze’s film HER. Now messaging app KIK has launched a Bot Shop. Bots are turning mainstream as the Guardian explains to its readers what a chatbot is and how to use it, “bots are the new apps” says Microsoft, and The Economist proclaims Bots, the next frontier. So “WTF is all the fuss about chatbots” as Digiday asked recently?  

You can view bots from at least four different perspectives; as an operating system (a new platform), as an interface (compared to web and apps), as a product (the user and consumer perspective) or as a channel (the enterprise and advertiser perspective, wanting to reach consumers).

Bots as platforms

Microsoft’s CEO Satya Nadella calls bots the next big thing labelling it “the conversation as a platform”. What does that mean really? In essence, it’s about using human language, our speech or words to interact with systems, and that the systems themselves learn from our language, behaviour and input. Bots are closely linked to Artificial Intelligence (AI), as they both process natural language, while also continuously learning from the data to handle our commands better. Amazon has launched Alexa, a toolkit for building voice-enabled software, for example empowering the Uber app with voice. Amazon has even set up a $100 million dollar fund, the Alexa Fund, to “fuel voice technology innovation”. KIK’s platform offers developers resources to build their own chatbots, positioning it as “chat as a platform”. Big tech companies are now increasingly opening up their platforms for external developers. Expect Apple to let entrepreneurs come up with new features for Siri. Facebook already offers developers to integrate apps with Messenger, and has now launched its own bot platform at this week’s F8 conference. Facebook messenger is used by 900 million people, compared to KIK’s 275 million users. This leads us to the next perspective, that bots will offer users new ways to interact and new interfaces.

Bots as interfaces

Bots may represent a power shift in the tech industry. Interfaces have evolved from websites (clicking hyperlinks with a mouse), to apps (navigated with our finger on smartphone touchscreens) to bots (controlled by voice or touch screen, but also increasingly proactive as they learn from us and offer suggestions). Is this the end of the web and apps? No. Bots will not change the app paradigm, just as apps did not change the web paradigm. It will just change the look and feel of apps. No new media has ever completely killed an old media, their characteristics and user value have just become more distinct in a new media landscape. Just think how books, radio and LPs are still around. Bots represent a new interface. Since the bot in itself is not really an interface (it rather uses other interfaces), you might want to ask where the bots actually live. For example, bots reside inside messenger apps like KIK or Facebook, in search engines like Google or in hardware like Amazon Echo. If you think of it, “voice” is not an interface, but it becomes an interface if you can control a thing like a TV or software like an app with your voice. The brings us to the third perspective, bots as products.

Bots as products

We are all familiar with the old experiments in digital assistants on websites. A dialogue box pops up on the homepage, with an “avatar” asking us if he or she can help. It never really scaled, and there was usually a human behind the scenes. This is not a bot, and it was probably easier just to call customer support. Now the interaction takes place within popular established interfaces like Google, Facebook Messenger, WhatsApp and KIK, building on new platforms and capitalising on the vast amounts of data. To make it simple, there are (currently) two main kind of bots. “The Chatbot” and “The Assistant”.

The Chatbot

As KIK explains in it’s launch video for the Bot Shop, “Ok, this isn’t that complicated. Bots are like people, but they’re not actual people, they are computers… Instead of a person, it’s a bot… Not robots – bots! They are like super smart apps you can talk to.” KIK’s new store launched with three categories of bots: entertainment, playful and useful. Some of the first brands to offer chatbots on KIK’s Bot Shop are H&M, Sephora and The Weather Channel. Just like you add a friend, instead you add a brand. A couple of years ago Kakao, the messaging app from Korea, started to offer something similar where you could add a pop band and start chatting, but the bot (or the band) seldom replied. We tested H&M’s chatbot and the experience is (or is supposed to be) that of chatting with a helpful person guiding you to shop a wardrobe that fits your style. But the “conversation” is more like answering multiple-choice questions and getting ready-made answers. As the bots aggregate more data and user behaviour, they will likely become smarter and more human-like. It’s all about the data, and that’s why the data-leaders like Google, Facebook and Amazon are ahead in the race. Chatbots are also called messenger bots.

The Assistant

A more general purpose bot, the Assistant can either be voice controlled like Apple’s Siri or Amazon’s Echo (powered by Alexa), or push-based like Google Now sending you notifications. Facebook M is a text-based assistant that lives inside Facebook Messenger. Anyone who has tried to talk with Apple’s assistant Siri gets the idea of voice as an interface, although it does not always work that well. One challenge that Benedict Evans of Andreessen Horowitz outlines in a recent blog, is whether the bot has general AI, or works for something very specific and narrow, like managing boarding passes at airports. Again, Assistants thrive on data, and the more data they digest the better they will be at helping us in a personal way. This will probably open up new privacy issues, as our data will both make these digital assistants more useful for us, while we also are giving up control of our data by letting tech companies aggregate it. Our data is today being used to attract advertisers to bots, particularly in messenger apps, which leads to the fourth perspective; bots as a new channel.   

Bots as a new channel

For anyone working with online marketing, the focus is on new, well-performing digital channels that can reach millions of users in a qualitative way. During the last 10-15 years we have seen the emergence of AdWords, AdSense and Google’s advertising network, then came Facebook, YouTube, Linkedin, Twitter and other channels. One puzzling thing is that the biggest channel of them all – online messaging – has no well-functioning ad networks, ad formats or advertising at all. Consider WhatsApp, the online messaging app with 1 billion global users and a daily volume of over 40 billion messages (more the double the world’s daily text/sms volume), and no advertising. WhatsApp dwarfs Google’s “only” 4.2 billion daily searches or YouTube’s 8.8.billion daily videos views. It was not surprising when Facebook acquired WhatsApp for $22 billion. Messaging will soon reach 2 billion people globally, according to eMarketer.

What if this huge, untapped opportunity suddenly offered marketers a way to reach consumers in a smart way? It would quickly become the world’s largest mobile ad channel. Except for a few interesting advertising experiments in Asia, especially at Kakao and WeChat, online messaging has not yet been monetised by advertising. Why?

One obvious reason is that messaging is a private and intimate channel and users do not want their conversations “eavesdropped”, interrupted by ads, or their data shared with advertisers. Imagine if you’re chatting with a friend about, say, movies, and the chat is interrupted by a commercial message from Netflix, just because the system has identified you as a likely target. Users would soon revolt.

Another reason is that there is not yet a native ad format that fits the characteristics of online messaging, in the same way that Twitter (Promoted Tweets) and Facebook (Mobile News Feed Ads) have developed native formats for a new medium. Chatbots could change that, and are likely to become the way for marketers to engage with consumers in messaging. Or rather the other way round; how users chose to interact with their favourite brands. KIK now offers just that, letting you talk with brands, and brands talk back with you. Marketers are wary not to scare consumers away in this new channel, as they are trying to find the right approach to marketing inside messenger apps.

So what are the considerations for companies?

First, brands like H&M are now asking themselves if they should develop chatbots (inside the leading messenger apps) and digital assistants (powered by platforms like Alexa). For example, if you are a SaaS company or a hardware manufacturer, would like to add intelligent voice capabilities? If you’re a retailer, for example, what is your strategy for living inside online channels with your customers as a bot? The question is pretty much the same as when companies back in the 1990s asked whether they should set up a website, or were thinking of developing smartphone apps in 2008 when Apple App Store opened. New technology brings new considerations.

Second, for investors, the question is now how to invest in bots and voice technology, as a new sector opens up like Bitcoins a few years ago. Expect specialised early-stage bot funds to emerge. Some of these funds might even be run by actual humans rather than bots…