When Prem Natarajan, Capital One’s new chief scientist and head of enterprise AI, came on board in May — after five years as a VP at Amazon, leading the Alexa AI organization — it was because he was intrigued. What was in the DNA, he wondered, of one of the largest banks in the U.S. and one with a reputation for a strong technology focus, that could help it succeed in implementing generative AI and large language models (LLMs) in a responsible, thoughtful way?
“Capital One was emerging in so many conversations as a big, forward-leaning investor in technology that was one of the first major companies to go all in on the cloud,” Natarajan told VentureBeat in a recent interview. Capital One “offered me a great balance for [the] next phase [of my career] — to contribute using my expertise but to learn about the new challenges that lie at the intersection of [generative AI] and the new set of customer and business problems.”
Natarajan, who leads technology strategy, architecture and development for Capital One’s enterprise data, analytics and machine learning initiatives, said the generative AI opportunity for enterprises is substantial — but far more so for organizations that have already committed to a technology transformation.
“They’re the ones that will be at the forefront of this,” he said. “People who suddenly wake up and say oh, this is cool, they may not be in the best position to harness it as pervasively as we can.”
Natarajan spent his early career squarely in research, at Raytheon BBN — a company best known for its DARPA-sponsored research (DARPA is an agency of the U.S. Department of Defense, responsible for the development of emerging technologies for use by the military). After that came a long stint at the University of Southern California (where he still has a faculty appointment), serving as a vice dean of engineering and as the executive director of its Information Sciences Institute.
But then, he said, he starting noticing a change. For the first time, he explained, the center of gravity was steadily shifting, year over year, from academia to industry. “I realized that this is where a lot of the new advances in AI were going to happen, because there was a potent combination of a lot of data — from serving so many customers on search, social media and ecommerce — and compute,” he said. “I thought maybe I should go spend some time in industry — to take a peek into what’s going on.”
After a few years at Amazon, Natarajan said he kept thinking about the verticals that really shape people’s lives — like healthcare, education and, not surprisingly, finance. What he saw in Capital One, he explained, is the “kind of bank a technology company would build.”
“When I look at the size of the technology workforce here — 12,000-plus people — and I look at the quality of the people I’m interacting with, this is certainly a technology company, at least in some sense,” he explained.
But Capital One, of course, also operates as a bank, with all the regulatory and compliance considerations that are necessary. To tackle that potent combination of technology and risk/compliance, he said, the organization requires a new operating model that scales.
(Capital One executives will be speaking at VentureBeat’s upcoming Transform event on July 11 & 12 in San Francisco, which focuses on the power of generative AI. Natarajan is also serving as a member of the AI Innovation Awards committee.)
“When I talk about Capital One being ready [for generative AI], it’s not just that they have the artifacts or this expertise — in addition to the size of the investment, there’s also the maturity in how to operate the technology workforce that sets us apart,” he said. Moving to the cloud means totally re-architecting the data environment, he explained. “These are not small tasks, these are multi-year journeys,” he said. “We are so many years into what is a required part of the ML and AI journey.”
Of course, Capital One is a bank, first and foremost, albeit one that is technologically advanced. And Natarajan emphasized that regardless of the sector, “there is a deep imperative to operate all of this in a responsible, thoughtful way — even more so for an organization like us, that is more ready technologically than most.”
For the longest time, he said, AI was about testing — such as having the right benchmarks. But now, Capital One has to take an inclusive AI approach right from the design phases of its applications.
“So do we have diverse perspectives represented? Are we challenging ourselves to think about the different outcomes?” he asked. Banks, he pointed out, have had to think that through for other parts of their business processes for decades. Now, he believes that Capital One has a “natural strength” to bring in multi-dimensional thinking and examine the different ways issues could manifest, from the design and implementation phases to the testing and ongoing refinement and improvements.
“We’re building applications that should serve the maximum number of people in equally performant ways,” he said. “To me, that’s the essence of a responsible portfolio.” Others can put together something that works, he explained, but it is essential to think through the guardrails and safeguards.
Even a company like Capital One is going through a learning and experimenting phase with generative AI and LLMs, Natarajan cautioned. “Everybody acknowledges, across every industry, that they are learning,” he said. “Everybody is exploring.”
For Capital One, customer service is certainly an early application contender. “But even there, we have to go through the process to make sure it actually works,” he said. “How does it improve the employee or customer experience?”
Natarajan said his top priority at the moment is to continue building a “world-class” AI organization. “We have the framework, we already have a fair number of AI and ML people,” he said. “I want us to be the top destination for the top AI talent that is interested in these problems. I think that’s what will prepare us most for the future.”
He added that he is inspired by the company’s 100 million-plus customers. “How can this world-class organization that we build accelerate the delivery of new experiences, differentiated experiences that make everybody’s lives that much easier?” he asked. “Capital One already has a strong data and technology-oriented culture — but everything can be strengthened, especially as we introduce new disciplines.”
Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More
When Prem Natarajan, Capital One’s new chief scientist and head of enterprise AI, came on board in May — after five years as a VP at Amazon, leading the Alexa AI organization — it was because he was intrigued. What was in the DNA, he wondered, of one of the largest banks in the U.S. and one with a reputation for a strong technology focus, that could help it succeed in implementing generative AI and large language models (LLMs) in a responsible, thoughtful way?
“Capital One was emerging in so many conversations as a big, forward-leaning investor in technology that was one of the first major companies to go all in on the cloud,” Natarajan told VentureBeat in a recent interview. Capital One “offered me a great balance for [the] next phase [of my career] — to contribute using my expertise but to learn about the new challenges that lie at the intersection of [generative AI] and the new set of customer and business problems.”
Natarajan, who leads technology strategy, architecture and development for Capital One’s enterprise data, analytics and machine learning initiatives, said the generative AI opportunity for enterprises is substantial — but far more so for organizations that have already committed to a technology transformation.
“They’re the ones that will be at the forefront of this,” he said. “People who suddenly wake up and say oh, this is cool, they may not be in the best position to harness it as pervasively as we can.”
Event
Transform 2023
Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.
A shift from AI research to industry
Natarajan spent his early career squarely in research, at Raytheon BBN — a company best known for its DARPA-sponsored research (DARPA is an agency of the U.S. Department of Defense, responsible for the development of emerging technologies for use by the military). After that came a long stint at the University of Southern California (where he still has a faculty appointment), serving as a vice dean of engineering and as the executive director of its Information Sciences Institute.
But then, he said, he starting noticing a change. For the first time, he explained, the center of gravity was steadily shifting, year over year, from academia to industry. “I realized that this is where a lot of the new advances in AI were going to happen, because there was a potent combination of a lot of data — from serving so many customers on search, social media and ecommerce — and compute,” he said. “I thought maybe I should go spend some time in industry — to take a peek into what’s going on.”
Capital One is the ‘kind of bank a technology company would build’
After a few years at Amazon, Natarajan said he kept thinking about the verticals that really shape people’s lives — like healthcare, education and, not surprisingly, finance. What he saw in Capital One, he explained, is the “kind of bank a technology company would build.”
“When I look at the size of the technology workforce here — 12,000-plus people — and I look at the quality of the people I’m interacting with, this is certainly a technology company, at least in some sense,” he explained.
But Capital One, of course, also operates as a bank, with all the regulatory and compliance considerations that are necessary. To tackle that potent combination of technology and risk/compliance, he said, the organization requires a new operating model that scales.
(Capital One executives will be speaking at VentureBeat’s upcoming Transform event on July 11 & 12 in San Francisco, which focuses on the power of generative AI. Natarajan is also serving as a member of the AI Innovation Awards committee.)
“When I talk about Capital One being ready [for generative AI], it’s not just that they have the artifacts or this expertise — in addition to the size of the investment, there’s also the maturity in how to operate the technology workforce that sets us apart,” he said. Moving to the cloud means totally re-architecting the data environment, he explained. “These are not small tasks, these are multi-year journeys,” he said. “We are so many years into what is a required part of the ML and AI journey.”
Implementing AI at Capital One in a ‘responsible, thoughtful’ way
Of course, Capital One is a bank, first and foremost, albeit one that is technologically advanced. And Natarajan emphasized that regardless of the sector, “there is a deep imperative to operate all of this in a responsible, thoughtful way — even more so for an organization like us, that is more ready technologically than most.”
For the longest time, he said, AI was about testing — such as having the right benchmarks. But now, Capital One has to take an inclusive AI approach right from the design phases of its applications.
“So do we have diverse perspectives represented? Are we challenging ourselves to think about the different outcomes?” he asked. Banks, he pointed out, have had to think that through for other parts of their business processes for decades. Now, he believes that Capital One has a “natural strength” to bring in multi-dimensional thinking and examine the different ways issues could manifest, from the design and implementation phases to the testing and ongoing refinement and improvements.
“We’re building applications that should serve the maximum number of people in equally performant ways,” he said. “To me, that’s the essence of a responsible portfolio.” Others can put together something that works, he explained, but it is essential to think through the guardrails and safeguards.
A ‘learning phase’ for generative AI at Capital One
Even a company like Capital One is going through a learning and experimenting phase with generative AI and LLMs, Natarajan cautioned. “Everybody acknowledges, across every industry, that they are learning,” he said. “Everybody is exploring.”
For Capital One, customer service is certainly an early application contender. “But even there, we have to go through the process to make sure it actually works,” he said. “How does it improve the employee or customer experience?”
Natarajan said his top priority at the moment is to continue building a “world-class” AI organization. “We have the framework, we already have a fair number of AI and ML people,” he said. “I want us to be the top destination for the top AI talent that is interested in these problems. I think that’s what will prepare us most for the future.”
He added that he is inspired by the company’s 100 million-plus customers. “How can this world-class organization that we build accelerate the delivery of new experiences, differentiated experiences that make everybody’s lives that much easier?” he asked. “Capital One already has a strong data and technology-oriented culture — but everything can be strengthened, especially as we introduce new disciplines.”
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.
Author: Sharon Goldman
Source: Venturebeat
