The Bot Chronicles SO1 E01: crashcourse
Hi again and welcome to the first chapter, Episode 1 of the Bot Chronicles series. I feel we got off a bit on the wrong foot in our initial article of this series. We jumped in head first and full of enthusiasm about this new bot technology, but we didn’t really check in with you guys. It’s all good and well if we know what we are talking about, but let’s keep you in the loop too, shall we? So let’s fill you in a bit more.
What are bots?
So first things first: what are bots?
I’m sure you’ll find tons of info on this very site and all over Google on this subject, but since you’re here, we might as well do due diligence and give you our spin on this new technology.
A “bot” is a piece of software that is designed to perform actions on your behalf. In general these are actions you perform yourself on a regular basis. Actions like setting a reminder for something, adding an appointment to your calendar, add an item to your grocery list, dial in the water/electricity/gas usage to your utilities company, etc.
These actions are fairly basic. We decided to use bots here at The Reference to figure out how it goes ‘in real life’. We are now using it mainly within Slack, our collaboration tool.
An example of bots within the company is related to our standup meetings.
Although the Agile Manifesto speaks of “Individuals and interactions over processes and tools”, we started playing with Geekbot, a bot that lives in Slack. We wanted to try out alternatives for our standup meetings. For those amongst you who are not familiar with a standup, it’s a 15 minute “meeting” where every member of a team says what they did the day before, what they’ll be doing today, and if there are any issues blocking them. Since we often have team members working in different locations throughout the company, we decided to give Geekbot a try. Geekbot notifies every team member from the moment they come online in Slack and posts three questions mentioned above. Every team member answers these questions when they see fit (some of them start very early, others start somewhat later). Geekbot kindly awaits their responses. When all members have done their thing, Geekbot summarizes everything and sends out a standup report containing a handy overview. This way, we can focus primarily on the blocking issues.
All in all this is basic stuff. But more and more we see these bots turn up in places where they interact with us on a more “conversational” manner: we ask it questions, or give orders. The bots determine the context and intent of our questions and responses and replies and performs as requested, but also provides us the feedback in that same conversational manner. Since conversations can be about anything, you start seeing bots in just about every domain there is. From marketing to customer services and content distribution over to full blown e-commerce solutions. Want to check out some of these? Botlist allows you to quickly filter based on platform and domain.
Where do bots live? — The Conversational UI
So just like a website is an interface for conversation and communication, a bot is one too. It’s just that, well, what is the interface here? What does it look like? Answering that question means diving into the realm of “Conversational UI”. #buzzwordalert! Any interface that allows for a “conversation” to take place IS the UI, hence “Conversational UI’s”.
The thing is, we meatbags communicate in many ways. There’s written communication, verbal communication, non-verbal communication (body language, tone of voice,…).
How then does a bot communicate with us? What is the conversational UI here? Well, basically it’s all of the communication types mentioned above.
So let’s imagine, in a David Attenborough kind of way, what this means for the “bot biotope”. Where might we find bots? Well, you’ll find them in those places where we humans communicate the most with each other. Care to guess where that is? You’re right of course: mobile phones.
And if we take it one step further. Which apps on those mobile phones do we use most often to communicate? Right again: messaging platforms. Looks like you’ve been reading up on bots on other sites, have you?
So there is our first bot biotope: messaging apps.
There are already plenty of bots out there on all of these platforms, but for now, bots in Facebooks Messenger are the most common ones.
But there are many different “strains” of bots. And like any other organisms, these strains might do better in a different biotope.
One specific strain even manages to cut its ties from phones or sites entirely. It broke free and invaded our cars and homes. These bots don’t leverage written communication. Instead they adhere to the spoken word.
In late 2014 Amazon launched the “Amazon Echo”, a voice-controlled intelligent personal assistant (called Alexa) in the form of a smart speaker with integrated microphones.
Besides Amazon related functionalities, the Echo also allows third parties to add new functionalities called “skills”. These skills provide added functionality like controlling your lights, or integrating with other devices like your scale or fitness trackers. It took only one year to go from 100 skills to over 5000 skills available for this device. Amazon expanded the product with the Amazon echo dot. All the smarts of the echo, without the speaker. The idea here is that you connect the dot to any speakerset you want.
In the coming years this bot biotope will perhaps become the most fought over. Amazon started a $100 million fund to invest in companies that will push these boundaries. Quite recently, Google jumped into the fray as well with their version of a true home assistant: Google Home. This platform was recently also opened to third party development.
Both devices are relatively low-cost and already a big hit in the US. This did not go unnoticed and as expected, this year’s CES provided a bunch of newcomers in this space as well.
- LG came out with the LG Hub (powered by Amazon’s Alexa)
- Lenovo jumped on board too with the Alexa (again) powered Harman Kardon speakers
- Even Mattel dabbles in the pool (oh dear)
- And of course we must not forget cars and the stray fridge!
2017 will be the proving ground for this technology.
Google’s “Google Home” implementation (Google Assistant) will tie into Google’s AI. AI integration into bots is a must have, but also tricky and difficult. It’s not something we at The Reference will develop ourselves. Just like most other agencies we’ll build on top of AI products created by Google, Microsoft, IBM,… We will need to take this into account later in this series when we discuss bot development platforms. I’m sure we’ll be spending some articles in this series to further explain the interaction between bots and AI, but for now, just take it from me that this home bot biotope will become huge (huuuge I tell ya)!
Messaging apps and your home aren’t the only biotopes. There’s the websites, and the business applications too, but as explained above, the growth isn’t situated here. We’ll have to follow the customer to where they communicate the most in a conversational manner.
So now what?
Does this mean websites and applications are doomed? Are bots in messaging apps on the smartphone and in our homes the holy grail?
Nah. Just like newspapers didn’t die when radio came along, and radio still survives after TV was born, and TV still is standing strong despite the Netflixes in the world, so will your sites and applications remain an important channel.
[queue video of a bot hatching out of an egg and read the next snippet in David Attenborough’s voice]
But they will have to share and co-exist with the newcomer in their biotope. A newcomer that might very well venture into biotopes they can’t follow it into.
The bot hatchling has a promising future ahead. But let’s not kid ourselves here. If I tell you that, in three months time since the launch of the bots platform on Facebook, there were already 11.000 bots in existence on messenger alone, you’d go “wawie”. But are they any good? And are they used? Do people like them?
Chances are, the answers is “some”, “some” and “some” again. There are many hurdles ahead in all kinds of domains. How do we create bot personalities people like to interact with? Is the AI good enough? Can we understand context and intent, and can we take in responses through Natural Language Processing? We’ll cover this and much more in future blogposts in this series, but at least now you know what we’re talking about here. On to Season 1 Episode 2 of the bot chronicles…
If we peaked your interest in bots for your own company, be sure to let us know.
Missed out on the Bot Chronicles first article? Read it here.