So here is the thing – I’ve built a chatbot. It’s not the cleverest bot I’ve seen. Still, this chatbot is a cool one. It communicates with clients, saving a lot of time of a tattoo artist. Here are some things I’ve understood while coding this chatbot. Hope you’ll find this post helpful!
1. You may build a chatbot for everything
When experts say a chatbot may be used for almost everything, they are right. This week I’ve been developing a chatbot that facilitates work of a tattoo artist. And I really loved it!
Chatbots may be useful in different spheres. This it the high time all progressive companies started building them. Believe me, when everyone understands the advantages of having a chatbot, this AI will be difficult to get. The demand will be higher. So the prices for chatbot development will also grow.
Though I’m interested in AI development and excited about all chatbots news, this was my first chatbot. It’s not perfect yet but I’m pretty proud of it. This is how the story begins. There is a famous tattoo artist who promotes works of other people via Instagram. He gets hundreds of messages from people who ask him about the process of promotion, its details, requirements, etc. In order not to waste so much time, the artist decided to organize the process of communication with clients with a help of a chatbot.
It’s a cool idea, isn’t it? The same may be done in a bunch of spheres: e-commerce, customer support, or any other type of business. Just tell me in what sphere you are working and I’ll provide you with reasons for using a chatbot on your website.
2. There are many ways to build a chatbot
When getting ready for my project, I’ve read a bunch of articles about different ways for writing a chatbot. I’ve thought that this is very difficult and I won’t do without someone’s help. I wasn’t right. It’s not that easy to code a chatbot, but it’s possible. I’ve found a long list of services you may find helpful if you want to build such type of AI for yourself. So how to code a bot? Just have a look at these services.
Rebot.me
This service allows people creating chatbots even if they don’t have programming skills. With Rebot.me users customize chatbots and teach them how to communicate.
Pandorabots
This service helps developers build and host website bots. Developers build chatbots using AIML and having API access to the platform that hosts bots.
ChatScript
This is an open source framework for all devs willing to create chatbots. ChatScript helps users create chatbots of the next generation. If you are looking for info on how to write a bot script, you’re welcome.
Microsoft Bot Platform
This platform also allows developers build bots. They do it using Node.js, Rest or .NET and Microsoft SDKs. Some of the most popular chatbots you may try out are Vision Bot, Bing News Bot, Murphy Bot, and CaptionBot.
Facebook Messenger Bots
Bots for Messenger is a tool allowing developers get access to 900 billion users. You may customize the welcome screen, send and receive API, get message templates. The most popular chatbots create with FB are as following: Spring, Hi Poncho, CNN.
So, these were the basic platforms that allow creating chatbots. If you are a developer like me, you may also use several services for connecting apps to voice recognition and enhancing AI: wit.ai, api.ai, Jasper, Clarifai, and Fancy Hands.
As for my chatbot, I’ve chosen another way. Not to exhaust you by detailed technical information or reveal all the secrets, I’ll describe the process of how to set up a bot in general.
For making the chatbot itself I used Facebook. This work also required a good knowledge of NodeJS so I’ve boosted my skills a bit. I’ve created a FB page and used Smooch’s messaging platform that creates channels for customer conversations. The database platform Heroku was also very useful. It’s the cloud platform that allows making, scaling, and delivering apps. At this stage, my chatbot started working.
3. My chatbot is clever, but not the cleverest one
So let’s have a look at how my chatbot works. This chatbot asks the person’s name and country. It also suggests one to submit a certain number of photos. After asking for the payment, it generates a PayPal link. All the information is put into a spreadsheet. When all the needed rows are filled and the payment via PayPal is made, the tattoo artist knows about it. Thus, he saves a lot of time he usually spends on constant answering messages.
The main goal of this chatbot is to collect all the needed information and generate a PayPal link. It doesn’t have to tell jokes, book tickets, or be funny with you. It has to collect the needed info and provide you with a link. So it doesn’t waste your time asking unnecessary questions or trying to be your friend.
If you would like to understand where this chatbot stands on a spectrum from “dumb bots” to “smart bots”, it’s pretty obvious. This chatbot doesn’t understand you if you try to chat with it about other things that aren’t mentioned in its dialog flow. This chatbot doesn’t learn. It doesn’t understand your voice. However, no one needs such features in this chatbot. It has its functions and it does the job correctly.
4. Chatbots are the future
While building this chatbot, I’ve found a lot of interesting information about the chatbot development. Many authors claim that chatbots create a new platform for communication. Taking into account the amount of time people spend chatting, they consider chatbots to be the future. That’s why so many companies are working on AI now.
There is a Google that has Google now. However, it doesn’t have many features. It will tell you needed info. That’s all. Though, Google is also planning to build AI. Allo is going to be a personal assistant that can analyze conversations, carry out some tasks for you while you’re chatting.
Microsoft is also working on making clever AI a reality. One of the last experiences you may remember is a Twitter chatbot Tay that became racist and rude less than in one day after its release. Taking into account its mimicking skills, it’s not that bad for Microsoft. Tay has been copying other Twitter users or just mimicking them. So we’d better think about the things we post on the Internet.
The experiment is over. Microsoft still has some work to do. It has another AI project that is more successful – Xiaoice. This chatbot has a memory. It’s popular among teens who chat with it every day, telling their feelings. Besides, Microsoft has got Cortana, a personal assistant that works on different devices and can carry out a lot of tasks.
Apple has got Siri. Siri understands language and responds to you via voice. Thus, it’s a conversational chatbot. Siri has many interesting features, though it’s not the AI you would like to talk to all the time. Amazon has got Alexa. It’s a really smart bot that can read off news, play songs or even call an Uber. There is also a Slack that sounds like a platform for chatbots.
As for Facebook, it seems to be a leader. Google and Microsoft need to work hard to beat Facebook Messenger that has more than 900 billion active users per month. Having analyzed all the statistics concerning the time people spend chatting via messaging apps and the number of online purchases they make, the reasons for Facebook to build a platform for chatbots become obvious.
All these examples show that big companies stake on chatbots. They suggest it to be something worth building. I completely agree. Besides, after I’ve created a chatbot, I understand how cool it is to become one of those people making AI real. This post is a bit long but now you know how to write a bot and what difficulties you may face. I’ve just wanted to share with you all the issues I find important and interesting. If you want to ask me anything, I’m ready to answer. Let’s chatbot!
[qzzr quiz=”239991″ width=”100%” height=”auto” redirect=”true” offset=”0″]
