Nosa Omoigui, CEO of Weave.AI, neuro-symbolic GenAI and intelligent agents that transform alpha decision making and risk analysis.
getty
In the traditional architecture of financial institutions, internal audit (IA) has long been viewed as the “rear-view mirror,” verifying compliance after risks materialize. That model is breaking down.
AI is changing this by making external signals, control evidence and peer activity continuously observable and testable at scale. In many failures, warning signs were visible well before losses materialized, but organizations lacked the ability to translate those signals into decisive action, resulting in capital strain, remediation costs, reputational damage and sustained board scrutiny.
As markets become more interconnected and opaque, driven by geopolitical volatility, third-party dependencies and rapid AI adoption, boards and executives are demanding independent, real-time and defensible assurance on where risk is emerging, not just where it has crystallized.
For chief audit officers (CAOs) and chief risk officers (CROs), this marks a fundamental shift. Internal audit is evolving from a retrospective control function into a horizontal assurance layer across the enterprise, providing the board with a continuously updated view of control effectiveness, risk exposure and systemic vulnerability.
Boards and audit committees now expect defensible, evidence-based insight into emerging risks and assurance gaps, not retrospective summaries. They want clarity on where assurance is insufficient and visibility into emergent risks.
Institutions that fail to evolve can risk creating a false sense of control, where assurance is incomplete, outdated or misaligned with external reality.
Assurance must shift from management assertions to independently verifiable, evidence-linked proof.
In a modern IA function, every finding must be:
• Traceable to underlying source signals
• Contextualized within the broader risk ecosystem
• Defensible under regulatory, audit and board scrutiny
IA is moving from validating what management believes to be true to establishing what is provably true, increasingly enabled by AI systems that connect unstructured disclosures, internal data and external signals into traceable evidence chains.
The collapse of Archegos Capital Management illustrates the consequences of failing to make this shift. Counterparties believed the risk was understood, yet total return swaps obscured the full scale of leverage and concentration. Without a consolidated, evidence-linked view to challenge those assumptions, exposures accumulated, contributing to losses exceeding $10 billion across major banks.
The lesson is clear: Risk is often visible before it becomes catastrophic, but only if evidence is translated into decisive action.
Evaluating controls within the four walls of the institution is no longer sufficient.
AI enables continuous benchmarking across peer disclosures, regulatory actions and market signals, allowing IA to surface coverage gaps, identify externally tested risks not assessed internally and challenge internal assumptions with independent evidence.
In this context, the absence of coverage is itself a signal. If peers are actively auditing a risk domain and an institution is not, that gap may indicate exposure, not strength.
Modern risk does not respect organizational or domain boundaries. Patterns such as third-party concentration risk, cyber vulnerabilities and governance breakdowns propagate across systems, counterparties and sectors.
IA is evolving from fragmented, domain-specific reviews to end-to-end, cross-domain assurance, connecting findings across silos, identifying systemic weaknesses and surfacing dependencies that amplify risk.
The result is an integrated, enterprise-wide view of assurance that reflects how risk actually manifests.
Traditional audit models suffer from data lag, making point-in-time reviews insufficient in a fast-moving risk landscape.
AI enables continuous, signal-driven assurance by ingesting internal and external data in near real time, detecting emerging risks earlier and dynamically recalibrating audit priorities as exposures evolve.
The Equifax data breach reflects the limits of static assurance. A known Apache Struts vulnerability remained unpatched, and controls failed to verify that remediation had actually occurred, leaving management assumptions untested and control effectiveness overstated.
Audit findings must link control gaps to financial, regulatory, operational and reputational consequences. For boards and audit committees, the question is not just what is wrong, but why it matters.
The Wells Fargo cross-selling scandal demonstrates the cost of failing to connect signals to systemic risk. Warning signs, including employee complaints and internal reports, were not escalated or synthesized into a clear risk narrative, delaying intervention and amplifying regulatory, financial and reputational consequences.
Across cases, including those cited above, three recurring failures emerge:
• Gaps in traceability, where institutions cannot demonstrate what was known and when
• Breaks in continuity, where episodic monitoring allows risk to accumulate
• A lack of challenge, where known issues are not tested against peer practice or forced into remediation
These reflect a systemic gap between available evidence and actionable assurance.
For CAOs and CROs, this shift requires operational change, enabled by AI-driven monitoring, prioritization and evidence-based challenge.
Three priorities stand out:
1. Establishing End-To-End Traceability: Ensure every risk signal and audit finding is traceable from source evidence through escalation and resolution.
2. Moving To Continuous Monitoring: Supplement periodic reviews with ongoing signal ingestion across internal data, third-party exposures and external events.
3. Embedding Peer-Relative Challenge: Calibrate assurance not only against internal standards, but against how comparable institutions identify and manage similar risks.
These are not incremental improvements. They represent a shift toward continuously updated, evidence-based, decision-forcing assurance increasingly expected by boards, regulators and stakeholders.
Internal audit is no longer a function that explains what went wrong. It is becoming the function that identifies where risk is emerging, where assurance is insufficient and what actions are required before loss materializes.
In an environment defined by complexity and interdependence, this shift is not optional. It separates institutions that anticipate risk from those that react to it.
Control failures rarely become existential because no one knew; they become existential because no one forced the issue early enough.
AI changes this dynamic by continuously surfacing risk signals, enabling IA to detect patterns and challenge gaps between assertions and evidence, shifting from passive reporting to active intervention.
Institutions that recognize this can redefine assurance as a strategic advantage. Those that do not may continue auditing the past while risk accumulates in the present.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。