Create a banking chatbot with FAQ discovery, anger detection and natural language understanding

In this developer journey, we will create a chatbot using Node.js and Watson Conversation. The Conversation flow will be enhanced by using Natural Language Understanding to identify entities and using Tone Analyzer to detect customer emotions. For FAQs, a call to the Discovery service will use passage retrieval to pull answers from a collection of documents.

When the reader has completed this journey, they will understand how to:

• Create a chatbot that converses via a web UI using Watson Conversation and Node.js
• Use Watson Discovery with passage retrieval to find answers in FAQ documents
• Use Watson Tone Analyzer to detect emotion in a conversation
• Identify entities with Watson Natural Language Understanding

Flow

1. The FAQ documents are added to the Discovery collection.
2. The user interacts with a chatbot via the app UI.
3. User input is processed with Tone Analyzer to detect anger. An anger score is added to the context.
4. User input is processed with Natural Language Understanding (NLU). The context is enriched with NLU-detected entities and keywords (e.g., a location).
5. The input and enriched context is sent to Conversation. Conversation recognizes intent, entities and dialog paths. It responds with a reply and/or action.
6. Optionally, a requested action is performed by the app. This may include one of the following:
   • Lookup additional information from bank services to append to the reply
   • Use Discovery to reply with an answer from the FAQ documents

With Watson

Want to take your Watson app to the next level? Looking to leverage Watson Brand assets? Join the With Watson program which provides exclusive brand, marketing, and tech resources to amplify and accelerate your Watson embedded commercial solution.

Included components

• IBM Watson Conversation: Build, test and deploy a bot or virtual agent across mobile devices, messaging platforms, or even on a physical robot.
• IBM Watson Discovery: A cognitive search and content analytics engine for applications to identify patterns, trends, and actionable insights.
• IBM Watson Natural Language Understanding: Analyze text to extract meta-data from content such as concepts, entities, keywords, categories, sentiment, emotion, relations, semantic roles, using natural language understanding.
• IBM Watson Tone Analyzer: Uses linguistic analysis to detect communication tones in written text.

Featured technologies

• Node.js: An asynchronous event driven JavaScript runtime, designed to build scalable applications.

Watch the Video

Steps

Use the Deploy to Bluemix button OR create the services and run locally.

Deploy to Bluemix

1. Press the above Deploy to Bluemix button and then click on Deploy.

2. In Toolchains, click on Delivery Pipeline to watch while the app is deployed. Once deployed, the app can be viewed by clicking 'View app'.

3. To see the app and services created and configured for this journey, use the Bluemix dashboard. The app is named watson-banking-chatbot with a unique suffix. The following services are created and easily identified by the wbc- prefix:

   • wbc-conversation-service
   • wbc-discovery-service
   • wbc-natural-language-understanding-service
   • wbc-tone-analyzer-service

Run locally

NOTE: These steps are only needed when running locally instead of using the Deploy to Bluemix button.

1. Clone the repo
2. Create Watson services with IBM Bluemix
3. Import the Conversation workspace
4. Load the Discovery documents
5. Configure credentials
6. Run the application

1. Clone the repo

Clone the watson-banking-chatbot locally. In a terminal, run:

$ git clone https://github.com/IBM/watson-banking-chatbot

We’ll be using the file data/conversation/workspaces/banking.json and the folder data/conversation/workspaces/

2. Create Watson services with IBM Bluemix

Create the following services:

• Watson Conversation
• Watson Discovery
• Watson Tone Analyzer
• Watson Natural Language Understanding

3. Import the Conversation workspace

Launch the Watson Conversation tool. Use the import icon button on the right

Find the local version of data/conversation/workspaces/banking.json and select Import. Find the Workspace ID by clicking on the context menu of the new workspace and select View details. Save this ID for later.

Optionally, to view the conversation dialog select the workspace and choose the Dialog tab, here's a snippet of the dialog:

4. Load the Discovery documents

Launch the Watson Discovery tool. Create a new data collection and give the data collection a unique name.

Save the environment_id and collection_id for your .env file in the next step.

Under Add data to this collection use Drag and drop your documents here or browse from computer to seed the content with the five documents in data/discovery/docs.

5. Configure credentials

The credentials for Bluemix services (Conversation, Discovery, Tone Analyzer and Natural Language Understanding), can be found in the Services menu in Bluemix, by selecting the Service Credentials option for each service.

The other settings for Conversation and Discovery were collected during the earlier setup steps (DISCOVERY_COLLECTION_ID, DISCOVERY_ENVIRONMENT_ID and WORKSPACE_ID).

Copy the env.sample to .env.

$ cp env.sample .env

Edit the .env file with the necessary settings.

env.sample:

# Replace the credentials here with your own.
# Rename this file to .env before starting the app.

# Watson conversation
CONVERSATION_USERNAME=<add_conversation_username>
CONVERSATION_PASSWORD=<add_conversation_password>
WORKSPACE_ID=<add_conversation_workspace>

# Watson Discovery
DISCOVERY_USERNAME=<add_discovery_username>
DISCOVERY_PASSWORD=<add_discovery_password>
DISCOVERY_ENVIRONMENT_ID=<add_discovery_environment>
DISCOVERY_COLLECTION_ID=<add_discovery_collection>

# Watson Natural Language Understanding
NATURAL_LANGUAGE_UNDERSTANDING_USERNAME=<add_nlu_username>
NATURAL_LANGUAGE_UNDERSTANDING_PASSWORD=<add_nlu_password>

# Watson Tone Analyzer
TONE_ANALYZER_USERNAME=<add_tone_analyzer_username>
TONE_ANALYZER_PASSWORD=<add_tone_analyzer_password>

# Run locally on a non-default port (default is 3000)
# PORT=3000

6. Run the application

1. Install Node.js runtime or NPM.
2. Start the app by running npm install, followed by npm start.
3. Use the chatbot at localhost:3000.

Note: server host can be changed as required in server.js and PORT can be set in .env.

Sample output

Links

• Demo on Youtube
• Watson Node.js SDK
• Relevancy Training Demo Video
• Relevancy Training Demo Notebook

Troubleshooting

• Error: Environment {GUID} is still not active, retry once status is active

  This is common during the first run. The app tries to start before the Discovery environment is fully created. Allow a minute or two to pass. The environment should be usable on restart. If you used Deploy to Bluemix the restart should be automatic.

• Error: Only one free environent is allowed per organization

  To work with a free trial, a small free Discovery environment is created. If you already have a Discovery environment, this will fail. If you are not using Discovery, check for an old service thay you may want to delete. Otherwise use the .env DISCOVERY_ENVIRONMENT_ID to tell the app which environment you want it to use. A collection will be created in this environment using the default configuration.

License

Apache 2.0

Privacy Notice

If using the Deploy to Bluemix button some metrics are tracked, the following information is sent to a Deployment Tracker service on each deployment:

• Node.js package version
• Node.js repository URL
• Application Name (application_name)
• Application GUID (application_id)
• Application instance index number (instance_index)
• Space ID (space_id)
• Application Version (application_version)
• Application URIs (application_uris)
• Labels of bound services
• Number of instances for each bound service and associated plan information

This data is collected from the package.json file in the sample application and the VCAP_APPLICATION and VCAP_SERVICES environment variables in IBM Bluemix and other Cloud Foundry platforms. This data is used by IBM to track metrics around deployments of sample applications to IBM Bluemix to measure the usefulness of our examples, so that we can continuously improve the content we offer to you. Only deployments of sample applications that include code to ping the Deployment Tracker service will be tracked.

Disabling Deployment Tracking

To disable tracking, simply remove require("cf-deployment-tracker-client").track(); from the app.js file in the top level directory.