Skip to content

In this workshop, we are going to create a customer service Assistant for a bank. We are looking to automate some of the top questions that are reaching our agents via support chat on our site,in this scenario - people trying to transfer money to friends or family.

Notifications You must be signed in to change notification settings

IBMDeveloperMEA/Build-your-AI-Chatbot-with-Zero-Coding

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 

Repository files navigation

Build your AI infused Banking Chatbot using new Watsonx Assistant UI

In this workshop, we are going to create a customer service Assistant for a bank. We are looking to automate some of the top questions that are reaching our agents via support chat on our site,in this scenario - people trying to transfer money to friends or family.

We will Start by redirecting people contextually into the bank's application if they have a transfer to make under $10000

Prerequisites

Step 1: Sign-up/Login to IBM Cloud

1.1- If you are an existing user, click on this link to Login: http://ibm.biz/nocodebot

And if you are not, don't worry! We have got you covered! There are 3 steps to create your account on IBM Cloud:

1.2- Put your email and password.

1.3- You get a verification link with the registered email to verify your account.

1.4- Fill the personal information fields.

** Please make sure you select the country you are in when asked at any step of the registration process.

image

Step 2: Create an instance of Watsonx Assistance service

2.1 - Fron Dashboard search for Watsonx Assistant service in catalog

1

2.2 - Check the terms box and click on create. This will take you to a new window

2

2.3 - Click on Launch Watsonx Assistant to start building your chatbot

3

2.4 - Go to top right, select manage > Switch to new experience, it will prompt a confirmation message, click on switch.

140980993-13e3a96b-29f1-4e8d-9e4e-50694575c48f

Step 3: Create your first assistant

3.1- Name your assistant and click on create 6

3.2 - From the home window, select your first action, this will take you to a new window. 7

Step 4: Create your First action

4.1 - click on create a new action

8

4.2 - add "I want to transfer $100 to my friend" and click save

8-1

We are going to setup 3 steps for our action

4.3 - change the title of the action in the top left to something more consumable - "Transfer funds"

4.6 - Step 1: In our first step, enter "How much would you like to transfer?" and Under options, we can define the type of response we anticipate from the user. select currency and click on next step

8-2

4.7 - Step 2: In our second step, enter "When do you want to make this transfer?" select date

Note that the system can detect dates in plain language, not just formatted (e.g. tomorrow, two weeks from now, etc...)

10

4.7 - Step 3 (a): In our third step, add a simple confirmation message to make sure we've got everything right. add "Just confirming, you want to transfer $ "$step1" on "$step2?"

Notice how we can use the $ sign to call upon pieces of information the user has already provided in an earlier step - these are called variables

11

In this step, you can also define options for our customers from directly within the step itself - unlike most tools which require you to set up a separate "entity" Let's add yes and no under options, We can even add some synonyms for Yes and No yeah, ya, nah, nope. Also change the skip step setting to "always ask".

9914888687f9cbaed1444ecdbaf8cda81db0697b

4.8 - Step 3(b): let's define what happens if the user selects Yes or No. In this step, user selects YES.

To set this up, we will add a condition, which must be met for this step to be triggered

add "Great, let's finish the transfer directly in our app where we will gather your account information and the transfer destination. Click here to continue." 

Highlight "click here to continue" and make it a link. In the link, type the following: "dtebank.com/transfer?date=(step1)&amount=(step2)" You can, in fact, add variables to your URL, if your transactional engine requires it We must then indicate to the system that this step ends the action successfully

a690876349b2fcc13fcb7a89c85b58d50d5061da

4.9 - Step 3(c): let's define what happens if the user selects No.

Add "Whoops! Let's start over."

In this case, instead of ending the action, we'd want to re-prompt previous steps so the user can start over. Select all of them so that the user goes back to the beginning of the action. Select every step (1-4) to re-prompt

f69f5ff044209b769ae127327f5208fc53ab1807

Step 5: Preview

To test your chatbot, click on preview and try it out like below.

ef0cd7abef668e469c138f68978aa47e54a56c2c

Step 6: digressions

5.1 - Truly complex situations, such as when customers change the topic of a conversation entirely mid-flow, are called digressions.

In this example, let's plan for scenarios where your customer asks what the transfer limit is during the transfer flow. To do this, we simply need to build a new action with a single step in response Create new action > add example "What is the transfer limit?" then, Edit step 1 "The maximum daily transfer limit for all customers is $10,000.". And then: End the action

5a5a803e7754df7a232738bd4096c23d593f7335

Step 7: Test Digression in Preview

7.1 - To test your chatbot's digression capabilities, click on preview and try it out like below. Here you can notice how Watson Assistant jumps back to previous step after digression.

7297d2e493c47f0c553c97338eb389ad367a804a

Step 8: Integration

Now we will see how you can integrate your Watsonx Assistant as web-chat and with third party tools

Web Chat

8.1 - Open Navigation bar, select integration and select Web Chat. Keep the environment as draft and click confirm.

12

13

8.2 - Here you can customize your Chatbot as below

a18ec6db6011caecea6f64180c74e55c66b4a416

8.3 - To integrate your chatbot with your application, simply copy the code from Embed section and paste it in Header of your application's HTML file and your chatbot will be integrated. You can test it locally as well as with your live application.

WhatsApp with Twilio

8.4 - To integrate your Watsonx Assistant you will need a Twilio Account. Once you have your Twilio account visit Link and follow the steps.

Things you can Try Further:

Develop an AI-infused automation tool to convert business FAQs to Watsonx Assistant-ready chatbot input

About

In this workshop, we are going to create a customer service Assistant for a bank. We are looking to automate some of the top questions that are reaching our agents via support chat on our site,in this scenario - people trying to transfer money to friends or family.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published