Risk management must move from cleaning up past failures to shaping business strategy from the outset. A controls leader argues that integrated governance, risk and compliance, supported by centralized data, preventive controls, scenario analysis and carefully governed AI, can help financial institutions respond to increasingly interconnected threats.
Risk management must be embedded in business strategy from the outset rather than deployed after products, technologies and services have already created problems.
A director of controls at a financial risk management company argued that governance, risk and compliance functions need to break down organizational silos and build an integrated control framework capable of anticipating threats rather than continually responding to past failures.
Speaking at CeFPro’s recent Risk Americas event, the controls director said the pace and combination of technological, geopolitical and operational change had made traditional reactive approaches increasingly inadequate.
AI, blockchain, quantum computing, geopolitical conflict and other emerging risks are developing simultaneously, leaving risk managers to understand how multiple threats could affect the same organization.
For her, the answer begins with integrating governance, risk and compliance with business strategy.
“It’s not like which comes first,” she said. “They both have to come at the same time in the beginning together. It has to be integrated.”
That requires organizations to consider risk when designing a product or service rather than waiting for something to go wrong.
The director characterized the approach as a “premortem” rather than a postmortem, asking what could derail a strategy before committing resources and taking products to market.
Traditional organizational structures can work against that objective. Risk, compliance, audit and business functions frequently operate in separate silos, creating disagreements over accountability and ownership.
Questions about who owns risk appetite, policies, scenario analysis and risk assessments can consume time while preventing effective collaboration.
The controls director argued that the precise allocation of every responsibility matters less than ensuring appropriate functions participate in the process.
First-line businesses ultimately own and manage risks, while the second line provides challenge, but excluding risk owners when policies or risk appetite are developed can create friction and reduce the likelihood that frameworks will work in practice.
“When you get rid of the frictions and the silos, when you allow people to support each other and talk when it’s not territorial, then it gets better,” she said.
A unified GRC strategy can also reduce duplicated effort and improve management visibility.
Centralized repositories are particularly important because policies and information provide little value if employees cannot easily access them. The same principle applies to data.
The director warned that organizations rushing toward AI can overlook weaknesses in their underlying data infrastructure.
Teams cannot model risks effectively without access to the necessary information, while inconsistent terminology can prevent different functions from understanding one another.
Standardized taxonomies and common definitions therefore become essential components of an integrated framework.
Technology can simultaneously strengthen the control environment. AI could allow employees to interrogate internal policies rapidly, identify jurisdiction-specific requirements and support continuous monitoring.
Used appropriately, it could also help with risk identification, profiling and mitigation.
But the director rejected a technology-first approach that ignores controls. Organizations already struggling with infrastructure and technology debt must address those weaknesses before assuming AI will solve their governance problems.
Her preferred principle for emerging technology reverses a familiar risk maxim: “Verify and then trust.”
Preventive controls are equally important. Access management, segregation of duties and approval processes should be designed before incidents occur rather than introduced through root cause analysis after losses.
Continuous monitoring can then identify problems earlier, potentially preventing relatively small incidents from developing into multimillion-dollar losses.