Manual to Autonomous: How Agentic Data Monitoring Helps Avoid Costly Reporting Errors

When it comes to regulatory reporting, good intentions are not always enough. You can attempt to comply with every regulation and still be blindsided by unexpected issues. We see the consequences of these issues all too often in the financial industry: Large financial institutions handed significant fines for oversights they didn’t intend.

Many reporting mishaps have data quality failures at their root, failures that are compounded by ill-equipped legacy monitoring systems. Fortunately, there’s a straightforward solution. Implementing an autonomous data monitoring system takes your team from reactive to proactive, inspiring more trust in your data and fewer regulatory reporting risks.

An ever-evolving regulatory landscape

Whether you call it “Basel 3.1“, “Finalising post-crisis reforms“, or “Endgame“, this next evolution of Basel standards will affect regulatory reporting for every financial institution. The new requirements include an aggregate output floor and standardized approaches for credit and operational risk, all of which rely on accurate data.

This is just the latest in a long history of increasing regulatory scrutiny, both on the local and global scale. Higher scrutiny leads to stronger financial institutions, but it also increases the reporting workload. When every new requirement stacks on top of everyday concerns like generating call reports and environmental disclosures, data quality concerns can fall by the wayside.

The very real cost of reactive data quality practices

Reactive data qualities often contribute to the stress of regulatory reporting. Many data governance teams are stuck working with legacy data quality monitoring solutions, which require extensive setup and onerous upkeep. This time expense means teams are firefighting issues instead of proactively improving data quality.

When data issues slip through the cracks of legacy systems, they affect the accuracy of your regulatory reports, whether that means the availability of records or the numbers contained within. And these types of errors come at a very real cost.

In just the last few years, we’ve seen a slew of data management–related financial penalties levied by regulators against major financial institutions. In 2024, a tier-1 U.S. multinational bank was assessed $136 million in penalties for failing to correct poor data management practices. That same year, a leading global investment bank was fined $348 million for failing to provide data from 30 trading venues.

If these are the consequences of overlooking missing, inaccurate, or incomplete data during reporting, is your organization prepared to withstand regulatory scrutiny? Organizations with substantial compliance risk profiles exacerbate that risk when they rely on reactive data quality practices.

Revamp your data quality strategy to reduce opportunities for error

There is another way. You can join the many successful financial institutions that treat data as a high-priority asset, and sleep easier as a result. It just takes dedicated attention towards data quality policies and monitoring systems.

Don’t want to take our word for it? Here’s what the United States Office of the Comptroller of the Currency has to say in their Regulatory Reporting Handbook (emphasis ours):

“Data quality is fundamental to accurate regulatory reporting and is typically dependent on the data’s source, which begins when the data enter the bank. To reinforce the importance of sound data governance, banks typically implement data quality standards, provide training to the appropriate personnel, and assess adherence to policies and standards on an ongoing basis. To be reliable, data should be processed and compiled consistently and uniformly.” 

That’s a good general strategy, but it’s worth adding in a little more detail. When developing data quality standards:

• Establish clear policies for responsibility and standards: Who will be responsible for which tasks? What will be the standard process for bringing in new data, or for analyzing existing data? Where will information about data governance live?
• Set up broad observability monitoring for all data. This should tell you whether your data is available and whether the metadata structure is stable.
• Set up deeper monitoring for the most vital datasets like transaction tables, insurance claim records, and datasets used when making credit determinations. You want to check for missing or incomplete data, data freshness, and general anomalies.
• Make a plan for unstructured data like customer feedback documents and call transcripts. These need extra attention, because you often can’t work with them as systematically as you can a regular table.
• Consider “unknown unknowns.” Legacy monitoring relies on manual rules where you explicitly define the failure criteria.If you only look for the issues you know might come up, you’re missing a whole slew of potential problems.

A best-in-class data quality monitoring strategy will also democratize access and allow subject matter experts to participate in data governance efforts. These team members often have a better understanding of what makes a specific table tick, and what kinds of issues are particularly worrisome. Ideally, subject matter experts would have access to no-code or low-code checks to fill table-specific gaps in check coverage.

Think of it like teamwork during a medical visit. Your doctor is an accomplished medical professional, but you’re the subject matter expert for your own body. You can only figure out the best care plan by working together.

Leverage flexible, scalable monitoring with automated data quality solutions

As you build out your data quality strategy, consider an automated data quality monitoring system. Unlike legacy systems that require manual rules for each of your tables, automated solutions can monitor hundreds or thousands of tables with minimal individualized setup. This gives you peace of mind without the time investment of traditional data quality monitoring.

Not all data quality monitoring tools are built the same, though. If you’re in the market, seek out a system that has these key features:

• Coverage for the entire data stack, including unstructured data. Your monitoring system should provide coverage from raw data through to processed data. It should also be able to monitor any unstructured data, like PDFs and call logs, that touches reporting.
• Efficient table observability monitoring, plus useful automated data quality checks. Table observability monitoring should give you frequent updates about availability and metadata, and be efficient enough to deploy across all of your tables. Automated data quality checks should cover risk areas you didn’t foresee, and adapt to normal shifts over time (such as seasonal changes in customer behavior).
• A straightforward natural-language UI for more intuitive data analysis. A user-friendly UI empowers compliance team members and regulatory reporting managers to quickly double-check whether the underlying data in a report is sound.
• A library of integrations and deployment options. Your monitoring solution should fit into your team’s current data governance workflow. It’s easier to get teams on board with a data quality strategy when that strategy can integrate with the tools they already use.
• Flexible notification routing that can be customized per table or check. You want to put information in front of all of the people who need to see it, but also only the people who need to see it. Increasing real-time visibility for reporting specialists decreases the amount of time wasted working with flawed data.

Some more minor features that are particularly helpful for financial services organizations are:

• The ability to compare data across sources, like staging to production tables. This type of data transfer is a common source of typo errors that can impact reporting numbers.
• Visibility into metrics by segment over time, including automatically-generated segments. Your data quality monitoring tool should be able tell you whether the calculated risk profile for one demographic segment jumped more than for others. These metrics often form the foundation for complex analytics used in reports.

At the end of the day, an automated and robust monitoring system should enable you to rest easy about your data’s health. Anomalo was built from the ground up to make data quality monitoring more functional for organizations with significant complex data quality needs. We’ve helped countless data governance teams stop firefighting and focus on proactive improvements to data gathering, data processing, and reporting.

How Nationwide found stronger footing with their most important data

As one of the largest insurance and financial services companies in the world, Nationwide is no stranger to regulatory reporting. They work with around 13,000 databases across the enterprise, of which around 5,000 are in production. This was difficult to manage in a legacy system that required manual rule creation for every table.

By integrating Anomalo’s robust data quality monitoring into their data governance plan, Nationwide increased both baseline visibility and deep data quality insights, while decreasing the company’s reliance on manual rules.

At the heart of the new plan is a data quality policy that requires the most important assets to be cataloged in Nationwide’s data catalog and set up for data quality in Anomalo. According to Mike Randall, Director of Enterprise Data Governance at Nationwide, “That was the best day of my life in data governance, when they came up with that standard and policy.”

Learn more about how Anomalo can help

Anomalo supports major financial services companies across every segment of the industry, enabling teams to build better data governance policies and trust the data that goes into regulatory reports.

Learn more about Anomalo solutions for financial services institutions.