Commercial Lending: Digital Automation

Sponsored Content

The Art of the Possible

By Joe Ganzelli Sr. 

Amid bullish bank stock prices, actual C&I loan portfolio growth has been an elusive goal for many banks in recent years. Is automation a missing link to a more productive and profitable commercial lending department?

Many traditional commercial lenders believe that digital automation (DA) doesn’t apply to their world because of their detailed underwriting analysis and loan structuring, but these traditionalists are becoming rarer each year. DA is the secret sauce that will help banks automate manual processes traditionally performed by commercial lenders, credit analysts and loan assistants.

To date, bankers’ attempts to automate commercial lending processes have largely focused on loan origination workflows and implementing customer/ third-party portals. However, workflows only provide marginal efficiency gains, because lending staff still perform numerous manual tasks. Meanwhile, many commercial borrowers have abandoned portal usage because they can’t easily view their loan status, comfortably perform self-service functions or satisfactorily expedite loan closings.

Commercial lending processes present a host of automation opportunities beyond traditional origination workflows, with payoffs that include:

  • Cost reductions
  • Reduced cycle times
  • Enhanced competitiveness/market preserve
  • Delivery improvements
  • Improved loan data quality and integrity
  • Transformed loan origination processes

Automating five key tasks could mean changing the commercial lending productivity game.

1. Marketing analytics aimed at identifying potential commercial credit request needs.
Public information on geography, NAICS, revenue and collateral filings can be leveraged to focus marketing efforts. Existing customer data can be supplemented with public information to feed basic algorithms that can identify and prioritize existing/prospective credit needs.

2. Financial “spreading” of borrower financial statements.
Most institutions still spread financial statements via manual data entry. Many loan origination systems provide for OCR and document ingestion automation, or the import of data from accounting platforms such as QuickBooks. Rules-based logic can classify this data into desired formats for import into spreading systems, with only cursory reviews for validation.

3. Validation of due diligence items and vendor ordering.
Loan origination systems often permit the creation of user-configurable rules to generate pre-decision and pre-closing items to confirm entity/collateral due diligence items. Knowledge repositories, pattern recognition and unstructured data processing to automate vendor ordering, status assessment and evaluation are all readily achievable DA capabilities.

4. Generate credit underwriting recommendations and decisions.
Although automated decisioning is predominantly used in consumer lending, many FIs have been using it in small business lending for years. Cognitive automation can expand the power of automated decisioning, notably on annual renewals and reviews, using artificial intelligence, super data sets, predictive analytics and evidence-based learning.

5. Early problem loan identification.
DA can connect borrower behaviors to identify red flags on potential loans. Examples include overdraft frequency/amounts/patterns, new/increased credit utilization, past due financial reporting and significant score changes.

Although largely still in its infancy, the “art” of digital automation is gaining industry momentum and already generating tremendous benefits for financial institutions. Banks that enact material and fundamental change to commercial lending processes can realize significant gains in efficiency and, more importantly, portfolio growth.


Joe Ganzelli Sr. is senior director, Cornerstone Advisors, a banking and technology consulting firm based in Scottsdale, Ariz.

  • Sign Up

  • Categories

  • Archive

Software: Kryptronic eCommerce, Copyright 1999-2019 Kryptronic, Inc. Exec Time: 0.064392 Seconds Memory Usage: 3.799858 Megabytes