Bot epidemic is everywhere. It’s undeniable that chatbots are the new trend for messenger apps. But chatbots themselves are just the tip of the iceberg. In fact, the move to chatbots is a much more serious disruption of the marketplace than it appears at the first glance.
Chatbots are a natural way of communication
In the real world, we interact using natural language. You don’t need to press buttons or voice strict commands to find out how your friend is doing. You just say “hey, what’s up?”
Today, the creation of bot platforms and chatbots is the hottest trend in the development of messengers. It seems that in the near future messengers that don’t support chatbots will be doomed to failure.
With chatbots help you can order food, check the weather, find information about your favorite singer, conduct banking transactions… almost anything.
Communication with a chatbot is based on a set of commands and phrases which are implemented in the chat program. Modern bots are able to learn from users and become smarter and more capable over time. This is how they differ from channels and public chats, the content of which is created by people. Chatbots are the fruit of years of work in the field of AI.
Where do bots come from?
The start of the bot era may be considered to be all the way back in 1950. That’s when Alan Turing’s famous article “Computing Machinery and Intelligence,” about artificial intelligence, was published.
Turing proposed the so-called Turing test, which helps to distinguish man from machine when it comes to written communication.
The first chatbots were ELIZA (1966) and PARRY (1972). These early bots were used exclusively to simulate human language and didn’t complete any other tasks.
Bots reached another stage of development in the 2000’s, when bots like Minotaur by Radiohead appeared. Minotaur had more than 1 million users and sent over 65 million messages, but it died in less than a year, just as many bots did during that period.
For a time, it seemed like bots were dead for good. But within a decade, completely new bots began to emerge. And they became more popular. Why are bots so big right now?
We are accustomed to sending messages. First, there were physical letters, then email, SMS, and finally Facebook Messenger, WhatsApp, and Telegram. We use messaging apps every day. No wonder they are some of the most popular apps in the world.
The growing popularity of messengers and chatbots is most noticeable in China. The most popular messengers there – WeChat and Baidu – have turned into real personal assistants. As of 2016, WeChat has more than 800 million users, Baidu more than 650 million.
These apps can accomplish a variety of tasks, from exchanging media files to shopping. And all this happens without leaving the app.
Why do you need to seriously think about chat bot app development?
1. The appearance of chatbot messenger platforms
Microsoft’s Bot Framework for Skype bots, Telegram’s Bot Platform 2.0, and Facebook’s Bots for Messenger have all been released within the past year. These companies, as well as other IT firms, have been developing their own AI-powered messaging bot platforms. Since people already spend a lot of their time with the phones in messenger apps, it makes sense to add additional capabilities to the apps that people already love rather than forcing them to jump between different applications and websites.
A lot of companies are already taking advantage of the opportunity to integrate their own services into a messaging interface. There’s the Product Hunt Kitty bot for Slack Messenger, CNN’s bot for Facebook Messenger, and the Foursquare bot for Telegram.
According to Business Insider’s report, in the beginning of 2015, the number of monthly active users of messengers surpassed the number of users of social networks.
There are even some new experimental bot stores. But like all unofficial app stores, they don’t have large audiences yet.
[Source: Chatbots Magazine]
2. New types of apps emerge.
As new varieties of apps emerge, they can replace those that have come before.
Chatbot developers follow the trend and create new products that help you to perform a variety of tasks using natural language.
Apps with AI emerge here and there. And they become more and more popular proving that they might be the future of mobile development. In 2016, we’ve seen apps such as Penny, an app that helps you manage your finances; Lark, an app that helps you achieve your fitness goals; and Luka, a messenger built to provide interactions between bots and people.
All of these apps use natural language and AI to facilitate interactions between a user and an app.
3. Typical apps aren’t as hot as they used to be.
Apps of the type we’ve gotten accustomed to are going to disappear thanks to so-called “app fatigue.” App fatigue refers to the overloading of apps on people’s devices.
According to Statista, despite the growing number of smartphones, the number of apps that we use regularly stopped increasing three years ago.
Read also: How to survive the future of apps
Two types of chatbots
If you’re already convinced that chatbots are the future of mobile app development, here are two types of bots you can develop:
1. Bots implemented in their own apps (like Penny)
These bots work within apps that perform a particular function (search for places to stay or make purchases in an ecommerce app) to automate interactions between a user and an app. This approach is suitable for businesses that already have many users.
2. Bots that function within messengers (like the CNN bot for Facebook Messenger)
These bots function in messengers that support bots. This type of bot makes more sense for businesses that don’t have a large audience yet. Plus, they’re cheaper than creating a standalone app.
Creating a chatbot not only gives your users a tool to solve a problem but also creates a platform with which you can communicate directly with your target audience.
How do chatbots work?
Chatbots complete tasks solely through text-based messaging. There are two ways that chatbots can understand text-based messages from users:
1. Strict commands
Chatbots that rely on strict sets of commands are very limited. They can only respond to perfectly formed commands that they are programmed to recognize. For example, Telegram’s music bot can search for music, but only if you use set commands, such as “/random” to get some tracks. If you try to chat with music bot using natural language, it’ll tell you it’s sorry, but it couldn’t find anything that matches your request.
2. Natural Language Processing (NLP)
Chatbots that rely on NLP are more intelligent and rely on machine learning. You don’t have to use strict commands to interact with them. Instead, you can speak to them like you would to a real person. You can text “Hey, what’s up in the world?” to get the latest news, or “Oh, I’m starving!” to get options for food delivery services.
Technology behind chatbots
You can build a chatbot that relies on fixed commands or machine learning with almost any programming language. Now let’s take a look at what technologies are used to create chatbots. We can divide these technologies into two kinds: APIs and Machine Learning technologies.
Let’s look at a simple example. Imagine you have an ecommerce site that sells shoes. There are two ways to interact with your site:
GUI (Graphical User Interface)
TUI (Textual User Interface)
The first option is that what we are accustomed to. A GUI has buttons, hamburger menus, tab bars, and so forth. So to buy a perfect pair of shoes you should perform the following set of actions:
Go to the website
Pick something you like
Put it in your shopping cart
Fill out the payment form
Wait for delivery
But chatbots work with Textual User Interfaces (TUI), meaning you can interact with a service using text commands. If you wanted to buy shoes, you could send a message to the appropriate bot telling it that you want to order shoes, and it could send you various options.
In order to implement a TUI and teach your bot to facilitate interaction between your base (which is an ecommerce site in our case) and a user, you have to integrate a particular API. The API connects your bot with the app you implemented the bot into and makes it possible to respond to a user’s requests and perform tasks.
Thus, from the technical side, the main requirement for putting a chatbot in an app is integrating its API.
We can use machine learning, as we previously mentioned, to make smarter chatbots that respond to more natural language commands. Machine learning technologies can allow bots to recognize speech and data, learn patterns of natural language, and interpret data based on previous interactions.
To make a chatbot better understand the queries every time a user makes them, layers of data are built one upon another. With time, these layers of data help chatbots learn from users.
What tools can you use to create your own chatbot?
There are tons of services that can help you build a chatbot, but they can be broken down into two general types:
1. Services that require some knowledge of programming
One of the most popular programming-based services is Pandorabots. It’s an open-source service that can help you create your own chatbot with the latest technologies.
With the help of Pandorabots Playground, you can build your bot in any programming language that can process multilingual input. Playground is a free IDE (Integrated Development Environment). You can find more information about building chatbots using Playground in this short guide.
After you’ve built your bot, you can integrate it into your app with AIaaS (Artificial Intelligence as a Service). AIaaS provides API access to their bot hosting platform, and SDKs that help developers integrate AI-powered components that they’ve created using Playground into apps. There are SDKs for several programming languages available on Github:
With the help of Pandorabots, you can also create your chatbot in AIML (Artificial Intelligence Markup Language). AIML is an XML (Extensible Markup Language) dialect which is used for writing natural language software. AIML is more flexible than other languages when it comes to building chatbots.
Pandorabots provides the base bot Rosie as a template for creating your own. Rosie is available on Github, and is a set of AIML and AIML 2.0 files which can form the base for any chatbot project.
There are no limits on your use of the Pandorabots platform. The most notable bot built with this service is Mitsuku – a conversational bot with millions of users that is available on the web as well as in Skype and Kik messenger.
api.ai is a service similar to Pandorabots that helps build natural language interactions for bots, application services, and devices
2. Services you don’t need to write a single line of code for
Chatfuel is one of the brightest examples of chatbot building services that require no coding. With the help of Chatfuel, you can build a bot for Facebook Messenger or Telegram for free.
Creating bots with Chatfuel is similar to building with LEGOs. You set rules step by step in a graphical user interface. You don’t need to write any code, and the Chatfuel team claims that your code will be maintained by them automatically.
With a Chatfuel bot you can send news, collect feedback, receive and answer questions and share content libraries — from GIFs to full business documents. You just have to plug in your own data streams.
The Chatfuel service is used by Uber, TechCrunch, National Geographic, Product Hunt, and other well-known companies.
Note: ManyChat for Facebook Messenger is another service that can be used without coding skills to create a bot for Facebook Messenger.
What does the future hold for chatbots?
In the near future, every kid will be able to create their own bot thanks to a wide variety of tools that let you build without a single line of code.
A lot of developers believe that chatbots can replace native applications altogether. Here are several reasons why that just might happen:
1. Creating, maintaining, and updating native applications costs a lot of money
2. With chatbots there’s no need for UI
3. Bots can learn from users.
4. There’s an overabundance of apps As a result, it is much harder for businesses to convince a potential user install an app.
We look forward to further news from the industry giants about the future of chatbots. It’s time to catch the wave and take advantage of these new technologies to achieve your business goals.