Wealth Management Chatbot

After you have the starter kit running (either locally or via toolchain on Bluemix) and had a look at the code, here are some things you can do next.

Train Conversation Service
Train the conversation service to understand a much wider range of queries so it is actually useful when discussing a financial portfolio
  1. Increase the number of intents (actions the user can do) and examples for each intent (to make the recognizer more flexible)
  2. Increase the number of entities (financial instruments the user can work with)
  3. Add new dialog nodes to give the chat bot the ability to respond to the new intents
  4. Add additional language support

Connect the bot to real user financial portfolios
  1. This will involve authenticating the user (probably through a mobile app or web page), but could also involve new dialog options where the bot itself prompts the user for credentials
  2. Query the user's financial data from a backend service (may require building a proxy in order to access the data through the firewall)
  3. Store the user's financial data in the Investment Portfolio services

Connect bot to a production user interface
Connect the bot to a production user interface (potentially a mobile app, web page, SMS, or even a voice interface)
  1. Secure the connection via HTTPS (direct to flow that involves getting certs for an app, setting a domain, etc)
  2. Sub-steps vary depending on target client

Improve Simulated Instrument Analytics
Improve use of simulated instrument analytics to support answering queries about additional scenarios (e.g. price of oil drops, or gold prices increase)
  1. This can be done by creating a large set of canned scenarios for the simulated instrument analytics service, but a better solution is to dynamically create scenarios (see below)
  2. Allow the user to create their own scenarios via the Predictive Market Scenarios service

Return up-to-date market values
Connect the bot to live financial data to allow it to return up-to-date market values for financial instruments
  1. Leverage Xignite collect stock quotes

Add graphing capabilities
Add graphing capabilities to allow the bot to include graphs in its answers to queries

Take the bot to production
  1. Deploy the application to multiple regions
  2. Increase the number of application instances to handle larger loads
  3. Leverage global load balancing
  4. Handling issues with connectivity to one or more required services gracefully
  5. Increase service plans to production levels
  6. Add monitoring
  7. Add api gateway to protect against denial of service attacks
  8. Improve performance of the API calls

You can also join the Fintech (Finance Technology) community at to join the discussion around using Bluemix in the Finance industry.