Latha Subramanian, SVP and Head of Data Engineering and Analytics at GM Financial, speaks with Rohit Choudhary, CEO of Acceldata, in a video interview about harnessing generative AI (GenAI) for data management and organizational efficiency, applications of AI, building scalable foundations, embracing agile methodologies, agentic AI, and the power of people.

Harnessing GenAI for data management and organizational efficiency

At the outset, Subramanian emphasizes the value of AI in streamlining critical data processes within IT. Focusing initially on developer productivity, she outlines several impactful ways AI can be leveraged to enhance data governance and management.

She highlights key functionalities AI brings to data management:

• Annotating lineage information
• Managing policy compliance
• Anonymizing data
• Augmenting data quality

These are foundational elements that improve overall data governance and ensure systems are not only efficient but also compliant.

In addition to day-to-day governance tasks, Subramanian points to the role of AI in driving infrastructure change. “It can also help you with providing accelerators for your data cloud migration if you want to increase velocity,” she says.

Broader organizational applications of AI

Beyond IT, Subramanian sees GenAI unlocking wide-ranging possibilities across the enterprise, such as customer experience. One notable initiative under consideration is a privacy chatbot. She says, “We’re also thinking about a privacy chatbot, which can look at all of the privacy guidelines and ensure that we can go through several pages and summarize the results.”

AI-driven solutions are not just theoretical, they are being actively developed and positioned for delivery and to help customers, says Subramanian.

Moving forward, she discusses the rapid pace at which technology, particularly GenAI, is evolving. In response to that, her team is taking a careful and strategic approach when deciding whether to buy solutions or build them in-house.

“We are being very deliberate about decisions on buy versus build,” Subramanian says, adding that part of this strategy includes maximizing the value of existing technologies. In terms of applying GenAI, she makes it clear that the organization is progressing with intention and caution.

Building a scalable foundation for innovation

Reflecting on the origins of the Data Cloud MVP, Subramanian states that the initiative began before her tenure, but its strategic drivers have remained clear and consistent. The need to enable innovation and improve productivity has been a primary goal that was difficult to achieve with on-premises infrastructure.

Highlighting the limitations of legacy systems, she says, “We cannot bring products, data, and solutions to the market faster if our platforms are not scalable enough.” Further, Subramanian says that while GenAI was not part of the equation back then, the organization still focused heavily on machine learning operations and the ability to accelerate innovation.

Over the past year, the team has strengthened the business case by quantifying tangible benefits:

• Labor savings
• Non-labor savings from retiring legacy platforms
• Cost efficiencies from existing data centers

Ditching waterfall for agile

Subramanian clearly states that traditional project methodologies will not suffice when it comes to cloud transformation. “If you are thinking about going for a data cloud modernization journey, you cannot operate in waterfall fashion,” she says. The shift to cloud requires a nimble, iterative approach supported by agile principles.

Then, Subramanian stresses the importance of laying a strong organizational foundation:

• Organizing the team around agile principles
• Establishing a sustainable, program-level framework
• Focusing on long-term scalability

Bringing customers along the journey

Another critical insight Subramanian shares is the need to engage both internal and external stakeholders early and consistently. “You have to make sure you are bringing your customers along on the journey, so help them adopt the new technologies that are on the cloud,” she explains.

To support this adoption, her team implemented several enablement strategies, including office hours to guide users through new tools and workforce upskilling for both customers and employees.

“We started on that journey early, and we also tried to create a culture of learning by creating communities, hackathons, and certifications and giving out recognition for certifications.”

A surge in innovation

According to Subramanian, there has been a remarkable appetite for innovation across the organization, particularly over the past year. Despite the heavy lifting involved in transitioning to the cloud, she observes that teams are energized by the possibilities of GenAI.

“We had a GenAI hackathon in our company, and we had close to 40 to 50 entries from the business unit on GenAI use cases.”

To build on this momentum, the organization plans to continue fostering innovation by recreating that same environment this year. Teams have already begun developing agentic AI frameworks to modernize systems and transition away from legacy platforms toward solutions like Databricks, which she describes as a very positive shift for the organization.

Subramanian also challenges the notion that AI might displace jobs, instead highlighting its empowering effect on the workforce.

Agentic AI, innovation, and the power of people

Sharing her experiences from a recent Gartner conference on data analytics, Subramanian says that organizations across industries are increasingly turning to AI not just for innovation’s sake, but to drive real, measurable improvements in operational efficiency.

Reflecting on what she saw at the conference, Subramanian remarks, “It’s refreshing to see so many companies using agentic AI to actually help improve organizational efficiency.” She cites practical examples, such as companies with auto service operations using AI to automate processes and increase workforce productivity.

Speaking of GM Financial, Subramanian says, “We have developed agentic AI internally; this is all homegrown, and we’re experimenting with it at the moment.” With a significant SAS footprint across the global organization and more than 150 scheduled jobs in production, the organization is now building a test bed to:

• Unpack existing SAS code
• Transition workloads to newer technologies like Databricks
• Help internal customers adopt and benefit from more modern platforms

While innovation is critical, Subramanian warns against focusing solely on the tools. She states that many major implementations fail not due to technical flaws, but because organizations neglect the vital interplay of people, process, and technology.

Subramanian firmly believes that change management matters and brings stakeholders along the journey. Concluding, she advises, “Innovate; empower leaders to innovate. Don’t hold on to any idea. No idea is a silly idea; fail fast, learn from your mistakes, and move on.”

CDO Magazine appreciates Latha Subramanian for sharing her insights with our global community.