(US & Canada) | Integrating AI/ML in MDM Can Achieve Higher Data Quality — Stibo Systems Director of Product Marketing and Solutions Strategy

Matthew Cawsey, Director of Product Marketing and Solutions Strategy at Stibo Systems, speaks with Lauren Maffeo, CDO Magazine Editorial Board Member, and Author, in a video interview about the role of technology in a data governance operating model, measuring the success of the master data governance initiative, and organizational challenges to successful implementation.

Stibo Systems offers a comprehensive platform to govern different data domains and drive insights, agility, and transformation.

At the onset of the discussion, Cawsey highlights technology as a critical part of the data governance operating model and says that it enforces the policies. Adding to that, he states that technology safeguards the organization, enabling it to define the business rules and policies within the Master Data Management system.

Further, technology plays a crucial role in providing tools for data integration, cleansing, and matching. He advises CDOs to look for MDM systems with governance features and data modeling.

Next, Cawsey says that having data quality and data profiling tools is critical to identifying and understanding data issues. Emphasizing collaboration across departments, he says that business process management and workflow management ensure data has the right approvals and the right set of eyes on it.

It is necessary to have role-based access control, especially around PII, and that is possible with technology, says Cawsey. He asserts that with the emerging role of AI and ML in data governance and privacy, organizations can achieve higher data quality, consistency, and efficiency by integrating AI and ML in MDM platforms.

Cawsey affirms that his organization has already started implementing the capabilities into MDM and MDG facilities, and those include algorithms for profiling, identifying data errors, and learning data cleansing.

One of the biggest challenges for data-driven organizations is duplicate detection, says Cawsey. He affirms that utilizing machines to comprehend patterns and algorithms significantly improves the quality and speed of duplicate detection in data governance.

Additionally, having a tool for anomaly detection, automatic classification, hierarchies of data classification, and categorization of data governance can result in consistent and repeatable results without constant human supervision.

However, Cawsey states that keeping humans in the loop is critical, but the focus should be on exception management and not routine tasks. Summing up, he says, technology is the mechanism by which organizations deploy data governance.

Moving forward, Cawsey shares that the master data management technology and software focus on core master data and providing data to other applications. The MDM software enables other applications to do their best work; it strives to generate the best value from them.

From both business and data governance perspectives, MDM is part of a broader ecosystem, says Cawsey. If an organization needs an enterprise-wide data catalog, MDM will be involved in cataloging master data attributes, he adds.

When asked how to measure the success of a master data governance initiative, Cawsey notes that measuring the success would be difficult for an organization that just started on data governance. Many factors create the success metric, which includes setting organizational intent, getting executive buy-in, and setting a budget for the initiative, among many others.

In addition, organizations must be able to demonstrate the business value of master data governance, says Cawsey. It is not just about data, but its real-life consequences. For instance, if customer data is not understood properly, a patient may end up getting the wrong dose of medicine.

Therefore, it is crucial to tie the data governance program to business objectives first, and then organizations need to focus on KPIs. Regarding KPIs, Cawsey prioritizes data quality, which involves tracking data accuracy, completeness, timeliness, consistency, and more, depending on the need. Again, he advises aligning KPIs with business objectives.

Continuing, Cawsey remarks that while there are no specific KPIs that fit all businesses, the key lies in understanding the strategic business objectives and what the roadblocks are. Inevitably, at some point, it boils down to data because all operations are data-driven.

Now, organizations must assess what is broken and how data governance can fix the breaks and improve business processes.

According to Cawsey, among some of the common mistakes that organizations make while implementing master data governance, one of them is inadequate stakeholder engagement. Thus, it is imperative to involve the employees in the journey and explain to them how this would personally benefit them.

It is upon the leaders to alleviate how data governance is perceived as a burden, and that can be done through education, stakeholder engagement, and executive support. This would also foster cultural change, says Cawsey, but bringing people along and changing their perception is still the biggest barrier to successful implementation.

Also, clear goal-setting is a must, along with departmental collaboration, rather than staying siloed. Cawsey advises organizational data leaders to bring in change management slowly but surely by understanding business needs, investing in the right talent, and adopting data governance policies.

In conclusion, he states that there may be tech, governance policies, and data advocates, but if there are no people on board, there will be no implementation.

CDO Magazine appreciates Matthew Cawsey for sharing his insights with our global community.