It was exciting speaking with Kamal Shah, co-founder of FintelligenX, a company at the forefront of realising AI's transformative potential in banking. With decades of experience delivering complex software platforms, Shah brings deep expertise to FintelligenX's mission of bridging the gap between modern tech stacks used by Big Tech and legacy systems in financial institutions. Through products like Hyper for code generation and XRay for intelligent banking, FintelligenX aims to enable autonomous banking with explainable, unbiased AI. As Shah shared his insights, it became clear that FintelligenX is playing a key role in ushering in the era of AI-first autonomous banking, empowering institutions to revolutionise customer experiences and operations while meeting regulatory needs.
Please walk us through your professional journey, and highlight the key milestones that led you to co-found FintelligenX.
In terms of experience, I have decades of experience in the software industry across geographies, working across multiple countries. I have been instrumental in delivering some very complex platforms and products across geographies. In terms of highlights, I would like to spend some time on maybe 3 or 4. The first one that comes to mind is Adhar, an initiative by the Government of India to provide digital identity to all the residents of India. I was a founding member of the Adhar team, responsible for the founding architecture, and wrote a lot of code for Adhar, which is today the world's largest biometric-based digital identity platform. The other flagship product I would say is Rubicon, a universal banking system that I was responsible for architecting and delivering. It is successfully running in more than 80 banks today across the world. The other two are working in partnership with the largest retail banks in the UK. The first one that comes to mind is a strong customer authentication platform, and the other one is an economic crime prevention platform that I'm still working on building with one of the largest retail banks. Now, how did FintelligenX come around? We have been seeing a quantum jump happening in software development. Organizations like Google, Meta, Amazon, have adopted modern tech stacks, cloud practices, and software practices. But the financial institutions are still looking at legacy tech stack, and they don't have access to this modern tech stack. So we thought about creating FintelligenX to bridge this gap and provide some groundbreaking software products and solutions to anyone who is interested in taking advantage of it.
Organizations like Google, Meta, Amazon, have adopted modern tech stacks, cloud practices, and software practices. But the financial institutions are still looking at legacy tech.
So FintelligenX is pioneering a vision for AI-powered autonomous banking. What sets your company apart from others in the market, and how are you revolutionising the financial landscape with your AI-driven approach?
We are on a mission to use AI to generate AI-powered software products and solutions. This is where our flagship product Hyper comes into play. Hyper is our AI-native code generation product that can generate financial-grade cloud infrastructure and application code. We are using Hyper to generate XRay, which is our flagship product for powering autonomous banking.
XRay is a suite of AI-powered software products and solutions. The way we see XRay revolutionising the financial landscape is through its unique agent architecture, where we want to use agents to replace and automate everything, replacing cumbersome manual processes.
Think of Hyper as an AI feature team that generates infrastructure code, cloud infrastructure code, and application code. A typical feature team in any organization comprises devops engineers, front-end engineers, and back-end engineers. Our vision is that Hyper should be able to generate code that a feature team can generate, but much faster, more reliably, and in a more consistent manner. We are utilising Hyper to generate code for XRay, which offers fully autonomous AI-powered software products and solutions.
Our vision is that Hyper should be able to generate code that a feature team can generate, but much faster, more reliably, and in a more consistent manner.
Why did you choose to target financial services and autonomous banking as the first application areas for Hyper?
Financial services have some of the hardest problems to crack in the industry because money cannot vanish, there are stringent regulatory requirements, and they have super legacy technology. We also have core domain expertise in this domain and wanted to stick to a domain that we are comfortable with and that is the hardest to solve.
Financial services have some of the hardest problems to crack in the industry because money cannot vanish. Made more complex with stringent security and regulatory demands.
Can you explain the relationship between your products Hyper and XRay, specifically how much of XRay was created using Hyper?
Today, the launched version of Hyper can generate 80% of cloud and DevOps code, 45% of back-end APIs and services including API specifications, 30% of UI and experience code, and 35% of AI training and operationalization code.
This is the percentage of code that Hyper generated for XRay. The biggest advantage we get with Hyper is that any change in specification, we can quickly embrace and Hyper is quickly able to adapt to those changes and generate the next version of the code that is easily replaceable.
In terms of contribution, I can say that Hyper is the biggest contributor to the quality, compliance, standardization, and harmonization of the code generated across all the different modules of XRay.
The greatest benefit of using Hyper is its agility. Whenever there's a change in specifications, Hyper swiftly adapts, generating an updated version of the code nearly instantly.
Diving deeper into XRay, can you provide examples of how it helps clients enhance their fraud prevention measures?
I'm going to expand slightly beyond fraud prevention measures and talk about XRay in general, because XRay is a suite of products and solutions to power an autonomous bank, and fraud prevention is just one aspect.
The way we have architected XRay is by utilising a unique agent architecture. The idea is that agents collaborate to execute complex processes. Some of these processes are powered by our own fine-tuned large language models. It not only performs regular banking processes like interest accrual and currency evaluation but also automates most manual processes.
The same architecture has been extended to our fraud detection products, a sub-suite of the larger XRay product suite, where it utilises the agent architecture and LLMs to drive intelligent fraud detection.
We have architected XRay by utilising a unique agent architecture. The idea is that agents self-organise and collaborate to solve complex processes by working as a team.
FintelligenX recently sponsored a roundtable discussion on combating authorised push payment fraud. Please share some key insights from that discussion and how you plan to contribute to this space.
Yes of course. First of all, I would like to personally thank all the participants of the roundtable discussion. I have worked with most of them, and I can vouch that they are some of the best brains in the industry to discuss this subject.
As you are aware, we are part of the digital sandbox at the FCA, and that gives us access to a lot of synthetic data. We are utilising the FCA's synthetic data and developing our own fine-tuned language models in this space. We plan to open-source these models to the financial institutions.
In terms of the upcoming FCA digital sandbox demo, we are going to showcase some fantastic demos of some of these experiments, the AI experiments that we are doing in our AI lab, which will be featured at the upcoming FCA event.
We are utilising the FCA's synthetic data and developing our own fine-tuned language models trap fraud. We plan to open-source these models to the financial institutions.
As the financial industry navigates the complexities of adopting AI, what do you see as the key success factors in the coming years, and how is FintelligenX positioned to support financial institutions on this journey?
In my view, artificial intelligence, especially with large language models, can shape the way customers interact with the banking system. AI has suddenly become a viable proposition to the common man.
Examples include providing a dynamic journey to a customer based on their past interactions, making smart product recommendations based on customer behaviour and usage, reliable interactions with chatbots, and safe onboarding journeys for new customers while keeping the bank safe from fraudsters.
These are some of the use cases where AI has a big role to play in automating the banking landscape. To achieve this vision, explainability and non-biased decisions from an AI perspective will be key drivers once you have full regulatory support.
FintelligenX is well-positioned to support financial institutions in their AI adoption journey. Our AI-powered products incorporate explainability from the ground up, prioritising transparency and fairness. By developing responsible AI solutions, we enable banks to harness the full potential of AI while navigating the complexities of compliance and ethics, ultimately helping them build trust with customers and regulators alike.
AI has suddenly become a viable proposition to the common man. Our AI-powered products incorporate explainability from the ground up, prioritising transparency and fairness.
Looking ahead, what excites you most about the future of AI-enabled software development in the banking sector, and how do you envision FintelligenX shaping this landscape?
The vision I see, where the whole world is heading, is a quantum jump happening in the software engineering space. In the future, we'll have AI feature teams where the code will be generated, but we'll have a human review cycle. You'll always require a human to review the generated code and make fine-tuned changes to it.
In terms of Hyper today, as I said, it is capable of generating code that requires a typical scrum feature team, comprising devops engineers, front-end engineers, back-end engineers, and quality assurance engineers. Hyper generates code that touches all these various engineering principles.
People and Agents should be able to collaborate and fine-tune the code easily. We expect that Developers will closely supervise machine generated code.
Can you share any upcoming projects, initiatives, or product roadmap items that our listeners would be interested to learn about?
We are now working on the next evolution of Hyper. I gave you some statistics earlier, where I said Hyper generates 80% of cloud code, 45% of APIs, etc. Our vision is that Hyper should be able to generate almost 90% of code across the full software engineering principle.
The next product is XRay. We are currently working on micro products in the intelligence space to look at fraud detection techniques. The next evolution will be bank-in-a-box for deposit products and lending products that will be fully Hyper-generated.
Hyper aims to create other AI products by chaining autonomous model selection, feature selection, training, validation and API based delivery.
Augment your Digital and Cloud Transformation with HYPER AI code generation agents. Develop Enterprise Grade Software Solutions at unprecedented speeds and high degree of standardisation, harmonisation and compliance.
At FintelligenX, we are a team of innovators and thinkers dedicated to making Financial Services more secure, more inclusive and affordable with AI.
Share this page