Turning Instant Apps in to Intelligent Instant Apps
How to build an Intelligent Instant App using Buzzy and IBM Watson
Abstract: A step by step guide on how you can create an intelligent instant app in minutes. I’ll cover adding an intelligent chat bot and using IBM Watson’s Personality Insights. At the end you will have your own customized intelligent app live on the cloud, available on phones, desktop & tablet. Setup a free Buzzy account and wow your friends and colleagues today!
What is an instant app?
An instant app is a cluster of important functionality and information that can be used on your phone, desktop or tablet, usually with a group of people. It could include important information, forms, the ability to chat and transact… instead of using multiple solutions and fracturing the conversation across email, chat and other systems, everything is in the one place — THE INSTANT APP.
Some examples of instant apps that Buzzy users have deployed include: coordinating an event, competition or club, creating and collating field reports from distributed sales teams, planning and coordinating group travel, rostering, collecting cash/payments from a group and more (see some examples here). An instant app can be created in minutes and is deployed and available to users immediately — no waiting around.
Here’s a quick video (1 min 20 sec) of how Love to Dream (inventor of 3-stage sleep system for babies) created a number of instant apps to help their business.
Making INTELLIGENT instant apps
Ok, so that’s great, but how do I make the app intelligent?
Ta-da!!! Enter IBM Watson running on IBM Bluemix (cloud platform).
IBM Watson on IBM Bluemix allows you to add intelligent capabilities to your apps in a matter of minutes. There are a bunch of IBM Watson capabilities like intelligent chat bots, visual recognition, IoT and personality insights that you can incorporate into your applications.
For the purposes of this article, we’ll focus on a couple of examples:
- Adding an intelligent chat bot to your application.
- Using IBM Watson Personality Insights to profile Job Candidates
You can leverage the Free Bluemix 30 Day Trial to get started with IBM Watson and create your sample Watson bots and services, which will be used for integrating with Buzzy.
1. Adding an intelligent chat bot to your application
Firstly, we’ll need an instant app. To get started quickly we’ll use an existing Buzzy application template. You can easily customize/tweak this.
Sign up for a FREE Buzzy account here www.buzzy.buzz to create your own apps, invite users and start using your chat bot.
As an alternative to the Job Candidates Template shown above, you can create your own instant app for anything you’d like. See this quick video on how to do that.
Creating an intelligent chat bot with IBM Watson is also pretty easy. Check out out this 7 min video. It obviously will take a bit longer to tune, teach and refine your bot… but you can get started really quickly.
Chat bots are useful for lots of things; for example, it can help you to scale level 1 and 2 support for you business, provide a conversational interface to an application (eg book a holiday), help you find some important data, access external services (eg weather) during a conversation… the list goes on.
So, assuming you’ve set up a basic chat bot on IBM’s Bluemix, all that’s left is to join the dots and allow users to access it… so the bot or “brain”, needs an (inter)face.. (mmm bad joke). That’s were Buzzy comes in. It takes approximately 1 minute to connect you IBM Watson chat bot with your Buzzy application. The bot will interact as if it’s a normal participant on a Buzz — ie you can send it private comments and it will respond to you as part of the conversation. You can then reply to keep the context and the conversation going.
Summarising the 3 simple steps to create a bot and connect to Buzzy (covered in video below too):
- Create your conversation services on Bluemix. Define your intents, entities and conversation structure. Get the credentials as you’ll need this to plug into Buzzy (watch this 7 min video).
- Create a new user in Buzzy that will represent your chat bot.
Hint: if you have a gmail address or gmail hosts your corporate email as per <email@example.com> you can add variations by using a <name> & “+” symbol… eg firstname.lastname@example.org can also be email@example.com, firstname.lastname@example.org etc.
Use these email addresses to create new Buzzy users to represent your bots. Once you have a Buzzy account for your bot, go into its profile, click ‘Add IBM Watson Credentials’ and enter the id/name, password and workspace ID of your bot.
- Now this Buzzy account/user is associated with your Watson Conversation Workspace (ie the bot) you can add it into any Buzz (application).
Test it out: Log in as a separate user who has access to your Buzz with the bot and send your bot a private message. You should see a response back.
Hint: Remember the Buzz needs to be published for the bot to kick into action. You can go back to your bot’s Watson workspace and train or tweak it to respond better.
That’s it for example 1! You can now chat via Buzzy with a Watson Conversation service/chat bot.
Remember you can re-use your chat bots across many Buzzes, simply by adding them as a user to the Buzz.
Hint: if you log in as your bot you can see the history of all the conversations it’s been having — you can then use this info to train your Watson Conversation bot. You can also reply to users manually as the bot.
2. Using IBM Watson Personality Insight to profile job candidates.
So we have a chat bot, which can chat with potential candidates and give them basic information on any questions they may ask.
We also have our job application app, which can store all the information (resumé/CV, ratings, comments etc) about candidates.
What if if we could automatically profile the applicants based on information they’ve written (eg on Twitter or other sources)? Surely that would make the selection of the right candidate for a team a lot easier? That’s where Watson Personality Insights comes in:
Personality Insights extracts personality characteristics based on how a person writes.
While Personality Insights is interesting for new hires, it may even be more interesting on building teams within organizations, to ensure they have the right mix of personalities for a specific team.
Setting up Personality Insights on Bluemix is pretty easy — click here for more info. In a nutshell, you create a Watson service. That’s it. Once you have your service, all you need are your “username” and “password” from the Service Credentials tab:
Now go to your Buzzy app and add these credentials… click the top left dropdown for you app and select “Properties”:
Then go to the “AI” (5th) tab and “Add IBM Watson Credentials”. Paste in your username/userID and password you just got from the credentials tab in the Bluemix console (described above):
Select the “type” as “IBM Watson Personality Insights” and “Done”(bottom right) to get out of there.
You’ve now successfully connected your Buzzy app to your Watson Personality Insights on Bluemix.
The next step is to associate fields in your app with Personality Insights, so it knows to use that text to generate the personality profile.
Go back into the Buzzy app “Properties” modal and select the “Fields” tab.
At the time of writing this article Buzzy recognizes two field types to associate with Personality Insights: a) “text” fields and b) “Twitter input”. You’ll also need to add an “IBM Watson Personality Viewer” field to your app, so you can visualize the output.
That’s it, you’re done!!! You now have an INTELLIGENT instant app that works on desktop, phone and tablet, can respond intelligently to comments via the chat bot, and can help with analysing the personality of job candidates.
You can make your app available to the people reviewing potential candidates; these users can then chat either to the group or privately. To add a user to the app all you need is their email address.
Pretty cool, huh???
We’d love to hear how you go. Don’t hesitate to email email@example.com if you run into any Buzzy related issues.