Can you build your own Personal Assistant chatbot? Yes, you can

Once when travelling, I sat next to the founder of a travel deals site who gave me an insight into how he significantly improved his company. What initially began as small talk turned into a conversation about how he’d formed his company with a friend, and most intriguingly, how they had added a new feature that not only improved customer interactions and had also significantly improved their business.

How? They had included a feature on the site that wasn’t particularly well known at the time — an accessible customer support window which would allow potential (but weary) customers to interact with the support team before making a purchase. This increased sales as potential customers received the validation, support and security they needed, before making big holiday purchases.

As we grow synonymously with the digital age, so does our search for instantaneous answers. We have a question, we Google it. If you have to ask a friend a question, you message them. If you don’t want to ask your friend, you ask Siri. The reality is, on a daily basis, we have questions that require answers. We’ve ditched our encyclopaedia sets because we now have the means to get those answers pronto. It’s on this thing I’m typing on right now, it’s also on the phone sitting beside my laptop. What that travel company did was allow visitors to get their answers immediately, without having to email customer support.

This initiative is great if you have 24 hour customer support. But if you don’t, what are your options?

One word: Chatbots.


Slack has them, and that’s certainly where I’ve seen them first, but it’s not the only place they exist. Facebook Messenger and Skype have chatbots as well. In fact, chatbots exist in many different hidden nooks and crannies of the web, continuing to grow in numbers.

We’ve seen them provide excellent customer support services for companies such as the Dutch airline KLM. They have been used to report, and raise awareness of social issues for organisations such as UNICEF. They have helped by becoming your personal lawyer and dispute your parking infringements for you. They also provide quick entertainment. The interactions are immediate, they’re always available… and if you’re a developer, you can make your own.

When Microsoft first purchased Skype, a few joked that this would be the start of SkyNet. Skype + Microsoft .Net Framework, get it, get it? I know, bad joke. Luckily they haven’t created killer machines; instead they have created the Microsoft Bot Framework. An open-source, bot builder Software Development Kit (SDK). Along with Azure Bot Service, this gives bot developers (you) the chance to build a bot and host it on their cloud platform — so you won’t have to worry about server and storage!

Now, creating chatbots mean, they’ll actually have to learn to… well, chat. That’s where Language Understanding Intelligent Service a.k.a LUIS comes in handy. It’s an A.I. built to teach your chatbot to understand language. The result is a chatting bot — but there’s a distinct difference in speaking to a bot and speaking to real person online, and that’s through tone, empathy, understanding…in other words, feelings. Well no need to worry about that either. There’s an API for that too named the Text Analytics API. Now you can build a chatbot, who understands you and your feelings. Nifty, right?

Now you’re probably ready to start looking into all of these tools, and assemble all the pieces together to create your first bot. Or, you could take our Mini Course on MS Bots. Within 6 short lessons we’ll guide you through the entire process of building a Skype bot that recommends movies using all the tools I’ve mentioned above.

What exactly will you be learning and how? Let’s dig a little deeper into these lessons.

  1. Getting started with the Microsoft Bot Frameworks
    We’ll show you how to get up and ready to create your first Skype bot using the Microsoft Bot Framework.
  2. Testing Skype Bots
    You’ll always find you run into bugs when developing anything. We’ll show you how to debug your Skype bot using the ngrok inspection tool, and test it using the Bot Framework Emulator.
  3. Natural Language Processing with LUIS
    No one likes a one-sided conversation, so we’ll create models in LUIS for our bot to understand and process natural language.
  4. Microsoft Cognitive Services with the Text Analytics API
    Apart from one-sided conversations, everyone wants to feel understood. We’ll integrate the Text Analytics API into our chat bot so our bot can chat accordingly with sentiment.
  5. Formatting Messages with Cards
    This is for the designer in you. We’ll show you (and your bot) how to format your Skype messages with images and buttons.
  6. The Azure Bot Service
    The Azure Bot Service is an industry first bot-as-a-service, service. We’ll teach you how to use this integrated online environment for this and all other bot development you may work on. There’s a bit to cover here, from the bot templates, online editor, integrated chat testing windows, and other configurations for further integrations.

This is only the start in your chatbot adventure. The real question is, after taking this quick course, what will your next chatbot do?

What will your chatbot do?