How to close the gaps: Scale effective model and data governance in insurance

Here’s a counterintuitive reality: model and data governance, when done right, is the opposite of a tax. It is an accelerator that allows insurance companies to deliver valuable innovation quickly, confidently, and responsibly, but getting it right can be a challenge.

When I sit down with insurance executives to discuss how they are leveraging models to generate value, the conversation brims with optimism and frustration. The value potential is huge, but often model and data governance can feel like an unfair tax. It’s a burden of paperwork, fragmented systems, and endless review cycles that slow progress to a crawl.

And I get it. In a highly regulated industry, protecting the customers and company from risk is incredibly important, but the process can feel like a treadmill that goes nowhere. I’ve witnessed the "governance gauntlet." One company I know had 47 people and many redundant processes involved in approving a single model. It's not surprising that model teams feel like governance is an add-on activity at the end of the build cycle, instead of appreciating the important role it occupies.

Three key governance challenges facing insurers

Insurance companies have historically trailed banking when it comes to the need for model and data governance because they lacked the intense regulatory pressure experienced by banks. This has left them with governance frameworks that have not evolved at the same velocity as their need to adopt and manage AI and machine learning models.

As insurance companies work to get their governance house in order, they are encountering challenges in three areas:

1. Cultural resistance: Model builder and governance teams have separate motivations and struggle to balance competing incentives.
2. Increasing regulation and model volume: New regulations such as the EU and Colorado AI acts are coming at a time when coding assistants allow model builders to build models much faster.
3. Fragmented technology and processes: Reliance on tools like spreadsheets, sharepoint, and email, paired with long established, bespoke processes, means model governance is difficult, expensive, and time-consuming to manage.

Not addressing these challenges leads to arbitrary variation. For example, different teams might use inconsistent assumptions for things like climate risks and replacement costs. The results can be catastrophic. HIH Insurance collapsed with losses estimated at $5.3B. The subsequent inquiry found that the primary reasons were under-pricing risk, insufficient reserves, and weak governance and risk management. (HIH Royal Commission*)

Beyond these issues, delays in deploying and refreshing models create risks and lowers model effectiveness to the tune of millions of dollars a year. The problem extends to reinsurers, too. If a reinsurer relies on a flawed model, it can distort the entire market, creating systemic risk and instability.

How to safely unlock value and accelerate AI governance insurance

What is Validation by Design?

To fix the bottleneck, insurers should start with a specific measurable goal. For example, reducing model validation timelines from six months to six weeks.

Getting there requires adopting "Validation by Design," which is a framework that modernizes three core pillars of an operating model:

• People: A cultural shift is required. Break down silos by embedding validators early in the design phase rather than waiting for a handoff after the models are built. Form cross-functional teams that align data scientists and validators under shared KPIs for compliant, and timely deployment. This ends the adversarial relationship and creates shared responsibility. A recent paper “Building Cross-Functional Collaboration Models Between Compliance, Risk, and Business Units in Finance” shows how collaboration between compliance, model risk, and business units enables faster issue resolution, stronger controls, and smoother product delivery. Additionally, the paper confirms that institutions that adopt shared governance structures and standardized communication frameworks achieve greater regulatory readiness, improved audit outcomes, and better alignment between risk management and growth objectives.
• Process: Move from slow, serial handoffs to fast, parallel workflows. Validators should begin work alongside developers, using a central library of standardized templates and "validation-ready" criteria. Companies should create standard model tiers so reviews are based on risk level. For example, an internal reporting model should not face the same scrutiny as a new, high-risk underwriting algorithm. Standards and frameworks minimize ambiguity. McKinsey states that automating key model validation processes can reduce cost by 20% to 30%. I’ve seen reductions as high as 60%.
• Technology: Replace the "glue and tape" of spreadsheets with a unified platform. This is the enabler that makes the People and Process changes work. When validators can stand up their testing environment in just a few clicks, time consuming administrative technical tasks are eliminated. A recent PwC report, confirms that a unified platform can yield dramatic reductions in model validation time and cost. Adopting automation and integrated end-to-end model risk management (MRM) platforms enhances governance and control while lowering costs and accelerating model development, validation, and deployment.

By unifying and optimizing these elements, insurers can eliminate validation backlogs, accelerate innovation, and unlock the high three-figure ROI that comes from getting models into production quickly.

Gain competitive advantage with innovative data governance and model validation

Adopting the "Validation by Design" framework is the key to moving model and data governance from being perceived as a tax to being embraced as a growth engine.

When your People, Process, and Technology are unified, the benefits are clear. You gain the agility to build new models quickly, optimize existing modes, and create new products all in a well-managed way. You don't just eliminate a roadblock; you unlock high three-figure ROI.

This is the new standard. Insurers that embrace this will gain a positive feedback loop of efficiency, trust, and innovation leading to sustainable competitive advantage. To hear more insights on governance in financial services, check out this eBook. Data science and AI leaders who are at the top of their game share proven strategies on how to scale governance, ROI, infrastructure, model validation, GenAI readiness, and more.

Source:
*HIH Royal Commission. “The Royal Commission into the Failure of HIH Insurance.”