AI Portfolio Risk & CECL Loan Review and Due Diligence

How Old National Bank increased loan review efficiency and capacity with AI-powered workflow intelligence

Bank asset size

$73 billion

Product

Loan Review Assistant

Results

~31% increase in loan reviewer efficiency

As Old National Bank's commercial portfolio grew, its loan review team faced increasing pressure to review larger and more complex credit relationships while maintaining consistency, quality, and regulatory readiness. By implementing Abrigo Loan Review Assistant, the bank increased reviewer efficiency by an average of 31% and expanded review capacity while maintaining strong governance and reviewer accountability.

Read the highlights below, or download the full case study to learn more about Old National Bank's experience.

Download the full case study

The challenge: Keeping pace with growth

When Old National Bank began expanding through acquisitions and organic growth, the bank sought a scalable way to increase review capacity while preserving its established review methodology and high standards for quality and governance.

Commercial loan review is inherently document-intensive. Before reviewers can reach conclusions about credit quality and risk, they must analyze credit memoranda, financial statements, annual reviews, collateral information, guarantor support, and a wide range of supporting documentation. As Old National Bank continued to grow, the volume and complexity of reviews increased as well. The challenge was to maintain the consistency and quality standards expected of an independent loan review function while supporting broader portfolio coverage with existing resources.

The bank had developed a structured review methodology designed to ensure that an independent reviewer examining the same information could arrive at the same Asset Quality Rating (AQR) conclusion. Applying those standards consistently across a growing review portfolio required significant expertise, time, and documentation effort. Old National Bank saw an opportunity to use AI to support that process without compromising reviewer judgment or accountability.

“Since adopting Loan Review Assistant across our team, we've seen firsthand how AI can drive both efficiency and consistency in credit review. It has enabled us to streamline manual processes, enhance the quality of our analysis, and better align our resources to where they add the most value.”

Sam Patton, Credit Review Innovation & Analytics Lead

The solution: AI acceleration that supports human judgment

Old National Bank partnered with Abrigo to implement Loan Review Assistant (LRA), an AI-powered workflow intelligence solution designed specifically for commercial loan review. The solution embeds AI directly into existing loan review workflows and combines document analysis, workflow automation, and institution-specific review guidance to support borrower-level reviews.

One of the areas where Old National Bank realized the greatest value was document analysis. Reviewers use LRA to analyze credit memoranda, annual reviews, financial statements, collateral documentation, guarantor information, and supporting credit files. The platform organizes information into a structured format aligned with the bank's review methodology and helps accelerate preparation of:

Purpose and repayment assessments
Financial analysis summaries
Guarantor evaluations
Risk and mitigant observations
Asset Quality Rating support
Review narratives and workpapers

A foundational principle of the implementation was that AI should augment reviewer expertise rather than replace it. Old National Bank implemented the following governance controls to ensure reviewers remained accountable for all conclusions and documentation:

AI-generated outputs are not examiner-visible
Reviewers validate AI-generated content
Analysts remain responsible for conclusions and documentation
AI-generated content serves as a starting point rather than a final answer

This human-in-the-loop approach helped establish trust, support adoption, and align with regulatory expectations for governance and oversight.

The result: Improved efficiency and reviewer productivity

Old National Bank has observed meaningful efficiency gains by reducing manual information gathering, document summarization, and preparation of draft review work product. Depending on the review type, the bank saw efficiency gains of approximately 15%-40%. These results reflect reviewer-reported improvements based on internal user feedback.

The strongest gains have been reported in new loan reviews, with reviewers reporting average efficiency improvements of approximately 31%, a typical range of 20% to 40%, and some reviewers estimating up to 50% faster completion for certain borrowers.

The impact of these efficiency gains is evident in the scale of document analysis the platform supports. In 2026, LRA supported the review of more than 430 borrowers representing over $8.7 billion in credit exposure, helping analysts accelerate review activities, navigate large credit files, and generate stronger starting points for risk assessment and documentation development.

Using LRA, Old National Bank now analyzes an average of 4,300 pages of credit documentation each month, including credit memoranda, financial statements, annual reviews, collateral documentation, and supporting credit files. This level of document analysis translates into nearly two full work weeks of analyst capacity gained each month, enabling reviewers to spend more time evaluating risk, supporting conclusions, and applying professional judgment.