In pursuit of greater efficiency and cost savings, banks are finding identifying new processes to automate with robot process automation (RPA) bots in the commercial lending space.

Commercial lending platform vendor AFS surveyed its banking clients about how they’re using RPA, which provided fodder for a Thursday webinar on the topic.

Brenda Alek and Bill Warren, directors at AFS, outlined 10 ways banks are deploying bots to create time efficiencies and cost savings. AFS did not release customer names during the presentation.

On average, the cost to implement a bot is a fraction of an employee’s salary at around 10 percent by some estimates, according to the AFS presentation. To reap that savings, Alek and Warren suggested banks should pick tasks that are:

• Repeatable;
• Predictable;
• Not overly complex;
• High volume; and
• Capable of being defined by a human.

Here are 10 ways commercial lending banks have used bots in the real world to automate or partially automate commercial lending:

1. Automate sending notices on syndicated loans

The first use case featured was a pilot AFS ran using its own system and a UiPath bot, called Digi, to automate notices on syndicated loans, which are loans offered by a group of lenders who provide funds to a single borrower. Handling syndicated notices can be time consuming to process manually given that banks can receive thousands of notices per day, typically sent as emails or faxes to member banks. The syndicated specialist gets the notice and looks for the borrower, amount and other data.

It’s time consuming, and transactions need to be processed in a timely manner, but it’s critical to get the transaction amount applied correctly since the dollars can be quite significant, said Alek. There can also be hundreds of institutions that are members, she added.

Digi cut out almost half of the effort of the transaction, automating about 80 percent of transactions, Alek said. It took about three weeks to get Digi the bot up and running, she added.

2. Prep tasks for portfolio managers and underwriters

Bots can prepare tasks for portfolio or relationship managers by searching a bank’s system for the needed loan documents, combining them into a PDF file and then notifying the portfolio manager or relationship manager that the file is ready for review and the next steps are needed. Essentially, the bot functions as an assistant that queues up work for employees.

“Another time saver for the front office is to move robotics to take information from the customer relationship system… and use robotics to key up work for the underwriters once the contact has become a viable prospect,” Warren said.

3. Produce customer letters

When financial information is required promptly for credit requirements, banks can let the bot generate the letter or request for more information. This involves automating a bot to review the database to know what information is required and when. This automation could apply to letters related to financial statements, insurance, continuations, or discontinuations, Warren said.

4. Automate checking KYC and OFAC watch lists

In another case study, one bank used a bot to extract information from electronic credit files and used it to check databases to see if the customer was on any watch list for Know Your Customer and Office of Foreign Assets Control compliance requirements. It’s ideal for robotics because it was “predictable and … a time-consuming process,” Warren said.

5. Validate payments and payoff inputs

Manual information validation can be a drain on payment processing efficiency. Bots can effectively validate these inputs. “In this situation, robotics simply needs to take the payment information capture, and then perform a workout and validation of an account and then post the proper time,” Warren explained.

In a related use case, another bank used bots to process when a cross-collateralized loan is paid off, changing the status and opening a workflow to queue up the release process for approval.

6. Extract information from PDF files during the loan onboarding process

A bot is deployed to “read” the PDF file and translate it into a workflow. A human picks up the workflow to compare what the bot puts into the field from the PDF file to ensure it’s put in correctly. This was particularly useful considering the volume of Payroll Protection Program loan applications, Alek said. Likewise, bots can be used to process deferrals for PPP loans.

7. Check legal documents for exceptions

Bots often work with optical character recognition (OCR) technology, combining an RPA bot with OCR to review legal documents for closing, for example. “At this bank, a robot searches for the documents and compares the documents to those it expects to find on a document checklist to make sure that all the documents have been received from the closing table,” Alek said. Bots can also check fields on documents for exceptions that need to be checked, she added.

8. Processing any rates that change frequently

For quick-changing rates, such as foreign exchange rates and index rates, bots can be equipped to communicate rate updates and pledge codes, Alek said.

9. Process late and auto charge changes

Account changes often received by the customer service department and manually handled, such as late charges or whether the borrower should come on or off automatic payment charges, can be handled using a robot, according to Alek. Bots can automate the process by reading account information and turning late charges or automatic payment charges on or off, she said.

10. Automating asset-based lending

Since collateral values change and affect the amount lenders can borrow on asset-based loans, timely updates to the collateral value are important. A bot can monitor the asset-based lending system to automatically update the collateral’s value in the loan system, Alek said.