(US & Canada) | Test and Learn Before Committing to GenAI Investments — Interac Corp., VP, Head of Data Analytics and Fraud

Palash Thakur, VP, Head of Data Analytics and Fraud at Interac Corp., speaks with Nazar Labunets, Product Marketing Manager at Ataccama, in a video interview about the organizational data governance program, the importance of data quality, the need for a data quality framework, building internal solutions with AI, and adopting GenAI.

Interac is a Canadian interbank network that links financial institutions and other enterprises to exchange electronic financial transactions.

Shedding light on the organizational data governance program, Thakur states that the program was intentionally designed to be adaptive. He adds that it is lightweight, adaptable, and fit for purpose while encompassing all facets of effective governance programs.

Further, Thakur states that the organization ensures that data governance becomes a force multiplier in its data transformation journey. It started with establishing data governance policies and standards, instituting an operating model, defining data domains, and identifying data stewards.

Adding to that, Thakur shares that various levels of data councils with executives, stewards, and SMEs were established as part of the program. Moreover, guardrails were put in place to mitigate risks, define the permissible use of data, and for enabling data-sharing frameworks and enterprise catalogs for better lineage and provenance.

Moving forward, Thakur mentions a soon-to-be-implemented automated data discovery and classification solution to secure sensitive information. In the case of AI models, he discusses putting a model governance framework in place so that the production models are reviewed, tested, explainable, well-documented, and bias-free.

Emphasizing the data quality aspect, Thakur states that It is a critical component of data governance. It is fundamental to identify its key dimensions and monitor and remediate the issues regularly.

According to Thakur, good data quality ensures the delivery of trusted insights and generates potential monetization opportunities. He says data governance is critical, but it takes time to explain its benefits to businesses and the broader organization.

When asked about navigating the uncertainty of data quality, Thakur emphasizes that with new products and services come new data sources, introducing organizations to new issues. While it is impossible to eliminate issues, having a solid data quality framework and process is crucial.

This framework should include monitoring key dimensions such as accuracy, data completeness, and timeliness, as well as a solid data quality issue management process, says Thakur.  

Education and literacy levels are also important factors in ensuring good data quality, he notes. Therefore, by continuously measuring and improving data quality, issues can be remediated, but achieving 100% issue-free data is likely unattainable.

Commenting on building internal solutions using AI and GenAI, Thakur refers to the primary organizational AI use case of fraud detection on payments. He stresses enhancing techniques and having a solid grasp of massive data sets.

Being well aware of the AI potential, the organization has recently developed an AI strategy along with a corporate strategy that identifies short-term, medium-term, and long-term opportunities. To make this effective, Thakur affirms speaking with the business areas, understanding their aspirations, and how they can align with the AI strategy to deliver on those.

Regarding GenAI, he advises organizations to first assess the use cases where the technology can provide benefits. Thakur suggests being an intelligent follower to reap a significant advantage.

Furthermore, he suggests organizations test and learn before committing to any GenAI investments. Thakur maintains that Interac is taking a similar approach to GenAI. He adds that the good thing about generative AI is that it can come up in existing tool sets, and the organization does not need to buy a separate platform.

In conclusion, Thakur states that while most of it is use case-driven, organizations must do the due diligence, let the AI capability mature, and consider leveraging it if it provides competitive advantages that drive aspirational goals.

CDO Magazine appreciates Palash Thakur for sharing his insights with our global community.