Skip to content

Content Library

Back to Articles

Article June 2, 2026

Building Model Risk Management for Growth: A View from the $10 Billion Threshold

Building Model Risk Management for Growth: A View from the $10 Billion Threshold

Crossing the $10 billion asset threshold is transforming model risk management from a basic compliance exercise into a strategic risk discipline, according to a senior model risk executive at a growing U.S. bank. The rise of AI, embedded automation, and r
  •        Crossing $10B in assets transforms bank MRM from simple compliance to complex strategic function
  •        Early stage banks manage few outsourced models with minimal staff and flat governance
  •        Post threshold requires larger MRM teams stronger governance formal reporting and audit grade tracking
  •        Validators shift to challenging models via benchmarking sensitivity testing and assumption review
  •        AI and embedded features blur model tool boundaries increasing validation volume and complexity
  •        Managing this requires adjusted risk thresholds automation and phased validations
  •        AI model opacity demands practical sensitivity testing to gauge reliability
  •        AI linked RPAs introduce new oversight needs requiring MRM IT collaboration
  •        Regulators expect evidence of active governance effective challenge and issue closure

Working in model risk management for smaller and growing banks has taught one senior risk executive that crossing the $10 billion asset threshold is not simply about size - it represents a complete operational transformation.

Having validated high-risk models across multiple institutions, the executive argued that growth fundamentally changes how banks must approach models, governance, validation, and enterprise risk management itself.

At smaller community banks, model risk programs are often relatively straightforward. Institutions may rely on a limited number of core models focused on asset-liability management, CECL, stress testing, and BSA/AML requirements.

These models are frequently outsourced to consultants, while oversight may sit with a single part-time risk manager supported by basic spreadsheets and relatively informal governance structures.

However, once institutions cross the $10 billion threshold, complexity rises rapidly.

“Suddenly, the models aren’t just simple regulatory checks,” the executive explained. “They become sophisticated engines requiring deeper testing, more assumptions, and specialized skills.”

As banks grow, outsourcing alone becomes insufficient. Validation work becomes more technical, governance structures become more formalized, and internal model risk teams typically expand significantly.

Reporting structures also evolve from simple committee updates into formal subcommittee governance with audit-grade issue tracking and transparency requirements.

At that stage, model risk management shifts from being a narrow compliance requirement into a core strategic risk discipline.

“The goal is not just to validate, but to challenge models effectively,” the executive said.

That challenge increasingly includes building challenger models, benchmarking assumptions against peers, applying sensitivity testing, and developing a much broader understanding of how models interact across the institution.

The executive argued that effective validators gain a uniquely holistic perspective across bank operations, allowing them to become strategic advisors rather than purely operational control functions.

Today, however, the greatest disruption facing model risk management comes from artificial intelligence and robotic process automation.

AI capabilities are increasingly being embedded into traditional banking models and operational systems - often without risk teams fully understanding the implications.

“Vendors are embedding AI features into familiar systems, sometimes without informing model users or risk teams,” the executive warned.

That development is blurring the historical distinction between a “model” and a “tool,” dramatically increasing the volume of systems potentially requiring oversight and validation.

Previously, institutions may have managed relatively small model inventories compared with operational tools. But as AI integrations accelerate, that balance is shifting sharply.

“More models mean more validations, more monitoring, and more people - unless MRM teams adjust how they operate,” the executive said.

According to the speaker, institutions now face three primary options for managing this surge in complexity.

Banks can adjust risk thresholds and accept that not every AI-enhanced feature requires full model oversight. They can invest more heavily in automation to manage growing validation workloads. Or they can divide validations into separate phases, reviewing foundational models independently from specific use cases.

Most institutions, the executive argued, will ultimately need some combination of all three approaches.

One of the most significant challenges remains the lack of transparency associated with many AI models.

Events you may be interested in

View all of our upcoming events

Sign Up to Our Newsletter

Sign up

Connect Plus

Unlock more with Connect Plus membership

This content is available to Connect Plus members.