Krish Swamy, Chief Data and Analytics Officer at Citizens Financial, speaks with Jonathan Steinert, Partner for Financial Services Technology at Guidehouse, about harnessing the power of centralized platforms and clean data, leveraging generative AI (GenAI) to improve both customer and employee experiences, GenAI use cases, and a human-centric approach. 

Scaling AI and ML in fraud detection

Swamy sheds light on how the evolution of internal data systems and platforms has significantly empowered data scientists to accelerate and scale their machine learning initiatives — particularly in areas like fraud detection. With the foundation of a well-structured data ecosystem, the traditional machine learning workflow becomes much more streamlined and efficient.

According to Swamy, the centralized machine learning platforms are at the heart of this transformation. These platforms allow data scientists to:

• Access curated data from the data marketplace
• Select and combine multiple datasets based on their specific problem statements
• Rapidly transform raw data into usable, model-ready formats

As Swamy explains, “It comes in a form which makes it easy for it to be combined to create the data sets that you want to start building the models on.”

This ease of access and integration drastically reduces the time spent on data wrangling, he adds. Instead, data scientists can focus on higher-value tasks like exploratory data analysis and deep understanding of the data.

This efficiency has translated directly into tangible results, says Swamy. “We rolled out a model in the fraud space late last year,” he shares. Additionally, “We are in the process of rolling out several more models in fraud, in areas like collections, credit trusts, and marketing this year.”

Some of these are enhancements to existing models, while many represent entirely new capabilities. “As a result of all of this, the number of models that we are building and adding to our inventory and repertoire is growing at a rapid clip,” he adds.

Swamy attributes this accelerated growth to a triad of enablers:

• Well-organized and accessible data
• Fit-for-purpose tools and centralized ML platforms
• Skilled, problem-oriented teams

Making human lives better with GenAI

Moving forward, Swamy delves deep into how GenAI is being integrated to improve both customer and employee experiences. From knowledge retrieval to summarization and complaint analysis, the focus, he says, is always on utility, control, and impact.

“We are building it on citizens’ data. That data is proprietary, it cannot go out,” Swamy emphasizes. This principle of data governance and security is a guiding factor behind the organization’s decision to develop many GenAI solutions in-house.

A prominent application of GenAI is in knowledge management, says Swamy, making internal and customer-facing information easier to access and consume. “How can we use AI to look at policy documents and how-to guides and bring that together in an easily consumable form for the users?” he assesses.

Whether it’s a contact center agent addressing a payment cancellation request or an employee learning about a new healthcare benefits program, AI-driven knowledge retrieval simplifies discovery, says Swamy.

For developers, the system also helps clarify deployment protocols such as the CICD and production code deployment policies that one needs to be aware of. He further explains that the team has created common frameworks to enable efficient data access and retrieval, regardless of whether the content is customer-facing or internal.

However, this initiative has surfaced an unexpected but valuable insight. Swamy shares, “In some cases, the documents are probably not clear enough, a problem that was maybe previously overlooked has now surfaced.”

Another major use case is summarization, especially in the context of customer service, he notes. For instance, there might be a requirement to create a summary of a 20-minute call between a customer and a contact center agent.

This summary can then be carried forward to future conversations, ensuring continuity and reducing the need for customers to repeat themselves.

Beyond enhancing service quality, Swamy highlights how these summaries can be leveraged at scale:  “We may be able to group 15,000 calls where the information is already available in a digital format so that they can self-serve.”

Thereafter, Swamy also discusses how GenAI is helping analyze feedback from a variety of sources, such as direct customer input, agent-recorded conversations, emails, and messages.

“Can we figure out what the root causes of complaints might be? Can we categorize them better?” he asks. The goal here is to derive actionable insights that influence product design or service experience — ultimately reducing customer dissatisfaction over time.

A human-centric approach

Across all these use cases, Swamy underscores a consistent agenda of enhancing human effectiveness.

“The common theme in all of this is how do you make the life of the human being better?” he maintains. “This human being here could be a customer, an employee, or a colleague.”

Whether through personalized assistance, improved service delivery, or smarter internal tools, the organization’s GenAI strategy is grounded in utility and empathy. “It’s focused on how we get them to be more effective at what they do and meet the ultimate customer’s needs a lot more effectively,” Swamy concludes.

CDO Magazine appreciates Krish Swamy for sharing his insights with our global community.