Kaushal Sheth is the first U.S. chief technology officer for Stuttgart, Germany-based digital banking service provider GFT Technologies, which sees the United States as a major growth market.  

The $1 trillion Deutsche Bank, digital banking service provider Salt and Hong Kong-based digital bank Mox, are among GFT’s clients. The company also works with Amazon Web Services, Google Cloud and Mastercard. 

Sheth, who has more than 30 years of experience in financial services at UBS, Misys, DeepMoney.ai, and more, told FinAi News the U.S. is a major AI innovation hub.

“We are quickly moving away from just pure legacy modernization or transformation to really building a new intelligence stack,” Sheth said. 

Sheth is looking to expand GFT’s AI capabilities, with a focus on AI solutions and agents guided by experiential intelligence. For financial institutions and manufacturers, these solutions will introduce a new level of recall and behaviors to operations, significantly increasing their competitive advantage and productivity, he said. 

Sheth sat down with FinAi News to discuss his new role and his approach to AI. What follows is an edited version of that conversation:  

FinAi News: What does it mean to build a new intelligence stack at GFT?  

Kaushal Sheth: In financial services, we have had a lot of innovation happening over the last 25 years, but now we are going into a new era of intelligence. 

For example, in factories, we had robots to automate different aspects of manufacturing and practically building technologies to automate complete product manufacturing life cycles. But there’s a difference between automation and AI. Automation was a very fixed rule-based algorithm, but intelligence is enabling the machines to learn and act and evolve from their environment. 

Similarly, in banking and financial services, we are building new models, new minds, from algorithms. We are shifting to cognitive intelligence. 

Those models allow and enable agents, and the agents can think and create new products and services for our customers, for better financial education and for better financial well-being for our customers.  

The shift from automation to cognitive intelligence is really the shift that we are seeing, and that has to be built ground up. We need new engineering, new architecture and new design. 

FinAi News: What products or processes will be powered by the intelligence stack? 

KS: GFT aims to power all processes through its intelligence stack, starting with KYC, fraud and data security issues.  

If we can enable more intelligent algorithms, more intelligent models, which can identify and augment the KYC process and evolve on their own as threats are evolving, that will be a great tool for FIs to have. 

Then we move to the next cycle of interacting with customers and understanding what their needs are. We want to enable FIs to provide customized financial products to their clients to aid them in their financial journey. 

FinAi News: What is GFT’s five-year AI-centric plan?  

KS: We see gen AI as the first step toward more complete and general intelligence. From gen AI, we will move to general intelligence or super intelligence [also known as artificial general intelligence]. 

We plan on not only having gen AI embedded into our solutions, but rather be the base on which all tools and platforms run. That gives us the ability to continuously evolve our services according to the changing needs without building a new infrastructure from scratch. 

The financial services industry will be one of the fastest adopters of super intelligence, because of the nature of the industry. 

It is a very objective domain where you can have very clear objective functions, unlike manufacturing, where there might be some more human aspects or cognition dependencies.  

FinAi News: Where does agentic AI fit in your roadmap?  

KS: We are building our own gen AI platform at GFT, called Wynxx. The platform acts as a sandbox providing agents with an environment where they can learn, and they can constantly and continuously evolve. 

Rather than building our own language model, what we are doing is building this framework, which brings together some of the best models and the best agentic capabilities. 

Models evolve from gen AI to more multimodal intelligence, and it will evolve into some of the other stronger, new experiential or large action models. 

We are working on partnerships with some very strong partners at Amazon Web Services and Google Cloud globally to build out our Wynxx platform and develop agentic AI models for banks to easily plug and play. 

Just like we shifted from models to agents very quickly, we will shift from agents to subject matter experts, because “agent” is a very generic term, or a very generic concept. The Wynxx platform will allow human talent to build subject matter experts, work along with them and train them in specialized job functions like CFA, financial analysts and risk auditors for increased efficiency and accuracy. 

Register here for the inaugural FinAi Banking Summit, taking place March 2-3 in Denver. View the full event agenda here.