Document labeling in automation

Problem statement

Every bank experiences problems with manual effort. In particular, employees need to process mortgage applications, work with invoices, and adhere to continually changing banking regulations by checking alerts. To address these problems, modern banks are implementing smart automation in collaboration with partners who have such expertise.

The Work.AI automation technology can automate manual effort, including paperwork operations. For instance, it allows digitizing documents and extracting necessary information from them for further processing.

While automation aims to free employees from routine and error-prone tasks, employees must first tell the machines what to do. These employees are the Subject Matter Experts (SME), on the bank's side, and the Data Analysts (DA), from the partner's side.

The bank and the partner begin the process of collaboration by explaining the bank's workflow to engineers and showing the partner's DAs sample documents they usually deal with. The SME plays a crucial role as a guide through the complexity of bank operations and document layouts. They also explain why specified fields of documents are important.

The main problem is that every bank has specific documentation. Banks have various departments that process different kinds of claims and applications, specific lists of clients who send different formats of documents, and business requirements for the materials and their processing.

Case study

Let's consider the example of an undisclosed global bank. It receives an overwhelming number of applications from clients every day. The overall workflow of the department working on applications is about 3,500 documents per day. As a result, the bank has 278 employees who process claims daily.

However, it is not the number of documents that make the work of the bank's employees so difficult. The amount is only the first problem. Other issues are even more dramatic and demanding. There are many types of documents, and they come from various partners, in different languages, and on different schedules. Specialists need to do the following:

1. Save the documents to separate folders.
2. Add details manually in the internal system.
3. Compare the details in another application.

If there are mismatches in the data, those should be added to the database. This process is fully manual, time-consuming, and prone to errors.

There are at least five problems here, listed in the table below.

Variety	There are 11 document types in total: 8 document types are processed in daily review and 3 more document types are processed monthly.
Priorities	Of these documents, types #1, 5 and 6 make up 90% of the workflow.
Connections	Document types 1 and 8 are interconnected: it's necessary to validate the data extracted from document 8 against document 1.
Different clients	There are 15 top vendors supplying invoices monthly.
Languages	Invoices come from 10 different countries in 5 languages (English, German, Japanese, Chinese, Spanish).

Solution

Our solution for the bank is to implement the Work.AI intelligent automation technology. To make it work, collaboration between SMEs and DAs is essential.

The main goal of the collaboration is to enable an Optical Character Recognition (OCR) engine, enhanced by ML algorithms, to recognize all documents in the Business Process. For this, the SMEs and DAs have to review every document type, discover the business logic, and find out how to label these documents. The reviewing process should account for the instructions from each bank's partner and the bank's requirements.

Hence, to know how to build an automation process to optimize the bank's operations, it is essential to study every document. The collaboration between DAs and SMEs is at the core of the labeling process and the first essential automation phase. The working environment is the WorkFusion's application called Workspace, which is a part of the Work.AI platform.

Next step

Now, it's time to learn about the DA and SME roles more specifically:

• How they can collaborate most effectively
• How to build a process from their collaboration in the Workspace application

We have a lot more for you!

This section is designed specifically for DAs and SMEs and covers only labeling practices. Though crucial, they are only a part of the automation process. If you want to study other aspects of automation, explore the Automation Academy courses where you can learn about high-level automation or the way to build an automation process from scratch.

If you are interested, try Automating the Business Process. If you want to learn more about automation in the banking industry, see the course Intelligent Automation for Banking.