IBM FiNancial Crimes Due Diligence 

Professional Work, 2018 | Focus - Product Development, User Interface | Time - 7 months

10% of the world’s top financial institutions spending $100+ million annually on KYC processes and the cost is rising.

Lack of skilled professionals capable enough to perform KYC procedures.

A bank, on an average, spends 24 days to complete the customer onboarding process and 12 percent of companies said they had changed banks as a result of KYC issues.

The Goal

How Might We Help companies run KYC compliance and due diligence faster and more efficiently by streamlining the customer onboarding process.

Research 

Working with Business Analysts to do primary research with vendors and our user group ( KYC analysts) to understand the business needs. As stock trade and financial crime was a new topic to me, I would often use design methodologies like mind mapping and personas to understand the archetypes and the challenges.   

Main Challenges

For User

Multiple logins with lack of brand language.

Just logging on to the system is time consuming due to multiple required logins. 

Generic login screen with no distinct brand representation. 

Manual personal 

Information Collection

Time consuming Information Aggregation Process

Collecting all the necessary information that needs to be collected from the customer is a long process that requires a lot of manual labour and time. 

Manually searching and going through various articles and documents in order to find information and news about a customer is a time-consuming process prone to errors. 

For Company

No single place to keep track of the amount times the product is being used.

The product was being sold to various external companies and IBM wanted to keep track of its usage amounts and understand how effectively the tool was working.

Suggested Solutions

Single login with clear brand  language.

Less time-consuming Customer Information Collection process

using Watson by leveraging natural language and image processing technology.

A system that provides financial crimes insight using Watson content Analytics by identifying relevant articles, providing contextual annotation and aids with the decision making process.

A royalty management system that can keep track of when the product is being used, for how long and if it was successful.

To-be Journey Map 

Journey Map to understand the flow of tasks that a KYC analyst needs to do inorder to create a consolidated report of a customer. 

• Login

• Search record / Create record

• Evaluate risk 

• Create report

Low-fidelity explorations

Login Screens

A variety of login screens were created to match the feel of the product and the following was chosen.

Others that were considered -

Customer Lookup - news analysis using watson content analytics [ mid - fidelity ]

The analyst is able to look through all new articles relating to the customer/company and the system is designed to give a quick summary and tag potential suspicious activity. The analyst is also able to help the system learn or correct itself by marking whether its tagging system is correct/incorrect. 

analysis.gif

1.

Quick look at basic customer information 

2.

List of relevant internal and externally sourced news articles about the customer, ranked according to relevance of content.

3.

Selected article shown with associated keywords found in it.

4.

A draggable vertical split bar to easily display all the necessary information on one screen without using multiple pages or modal windows. 

1.

analysis2.gif

A feedback loop system with a quick toggle feature for the analyst to train the system on whether it is classifying data correctly.

Onboarding Customer Information Collection

This platform was built to help KYC agents onboard customer [ individual or company ] easily. It also allowed for customer lookup which shows the past/ongoing accounts of the customer in question.  

• Quick and easy onboarding.

• Automatic checklist for maximum efficiency.

• Scan or drop images/documents

Information is easily scanned from uploaded documents using Image Processing and Natural Language processing systems. 

Easy Auto-fill

Completed customer document that shows chain of approval, in-order to avoid confusion about changes made and who the document is currently with. 

Chain of Approval

Image and Natural Language Processing Studio

This is a training platform that analysts use to train the program to read drivers license and other documents necessary for onboarding. 

Dashboard

Log into dashboard with various options and a gallery of documents/image library

Upload Document

The system immediatly reads the information that it can process. This is shown on the right panel.

1/2

Preview

Preview button was added because viewing the image in the middle takes a lot of processing time. 

1/2

Training

Manually click and drag over areas that are not recognized in order to train the program. Information is displayed on the right panel.

1/3

Complete training

Check the model data and press the complete button when done.

1/3

Royalty Management 

Backend program to help IBM keep track of royalty management for various companies using the product.

The graph was changed to a bar graph because testing showed that analysts found it easier to use it.

Working with design systems

Working with elements from the IBM design system while also creating new components. 

Success metrics

Eradicate  
 55% 
false positives and increase accuracy and efficiency
Reduce  
 50% 
of time taken to go through the KYC due diligence process.
Achieve  
 100% 
auditability and standardization within the product realm.

Reflection

This project was a very exciting and fast-paced one. As a designer, I had to constantly juggle wearing different hats. This project had me on my toes the whole time. I was constantly collaborating with developers, business analysts and project managers to update and re-evaluate every decision that was made. 

Developing products that serve millions of people gave me a thrill and a sense of fear. But testing again and again with the users and re-iterating accordingly made me confident in what we were putting out into the world. 

Working with multiple teams spread across various countries taught me valuable lessons about
efficient communication skills and honed my ability to present a good product story. Overall this was a great project that served as a wonderful last project in IBM before I left for my Graduate program. 

The End.

Aparna Marcelin . Portfolio . 2019