While artificial intelligence dominated the discourse at Davos, with CEOs foreseeing its future prominence, companies remain at different stages of exploring AI's potential. For many businesses, this inevitable shift goes beyond merely enhancing efficiency; it's more fundamentally about reshaping their business strategies and creating new opportunities.
Yet, this advancement brings its own set of challenges. Integrating AI into existing systems demands significant resource investment and skilled expertise. Ethical considerations, such as data privacy and algorithmic bias, add layers of complexity. Despite these hurdles, the potential rewards are vast, said a senior technology leader.
In an exclusive interview with TechObserver.in, Kyndryl's India CTO, Sreekrishnan Venkateswaran, said, “We are observing more opportunities than challenges with Generative AI. This technology is enabling us to address and solve customer problems that were previously hard to reach. It is narrowing the knowledge gap between different entities, making information and skills more accessible.”
In view of growing impetus on Digital Public Infrastructure, how do you view the evolving digital ecosystem in India and Kyndryl's position within it?
Kyndryl values India for its unique market challenges and opportunities. The country's emerging market status introduces complex engineering problems and price sensitivities that stimulate innovation. This has led numerous global companies to establish R&D centres in India, recognising its potential for creative problem-solving.
Consider the example of a widely used mobile application like the COVID tracking platform in India. It's a testament to the country's digital infrastructure, having managed over 2 billion doses. The transaction volumes varied significantly, from about 20 per second during off-peak times to over 20,000 at peak, and even spikes to around 1,50,000 per second during major government announcements. Such fluctuations demand innovative solutions that often involve cloud-based technologies and advanced application design, including adopting event-driven programming. The app had to adapt to unpredictable workload variations with serverless architectures, alongside managing the predictable components differently.
The experience and innovation developed in response to India's challenges in terms of scale and diversity have wider implications. They provide valuable insights and methodologies that are applicable across various industries globally. Thus, the solutions and innovations crafted in the unique context of the Indian market have the potential to influence and shape global trends.
Considering the trends you have mentioned, how focused is Kyndryl on R&D? What contributions are you making in India that apply locally and globally
Intellectual Property (IP) is critically important to our operations. We place a strong emphasis on managing IP, re-utilising assets, and exploring innovative solutions. For instance, consider our work with prominent banks, a key sector among our clients. The recent rise in API banking in India has drastically increased the volume of transactions. A typical large bank we collaborate with handles over half a million transactions each hour, equating to about 10,000 transactions per minute. Therefore, a single minute of downtime can result in significant customer dissatisfaction, a situation more acute in the India context than in many other markets.
Given the massive scaling that India calls for, the requirements for resilience, digital immunity, and high availability are especially stringent. These standards are frequently highlighted in Requests for Proposals (RFPs) and included in contracts, sometimes with penalties for non-compliance. Meeting these stringent requirements necessitates sophisticated engineering design. For example, achieving only one minute of downtime in such a demanding environment equates to aiming for over 99.999% availability. This degree of engineering precision, while developed for specific scenarios, has a wide range of applications across our global clientele.
Our strategy focuses on the efficient use of existing IP and assets. We aim to leverage what already exists, adapting and repurposing solutions and patterns as needed. We maintain a collection of these resources, bringing them into each new project as a base. This approach allows us to begin with a solid foundation, customising solutions to meet each client's unique requirements. Overall, our approach is to employ a comprehensive and adaptable strategy in IP and asset utilisation, ensuring high-quality, tailored solutions for a variety of market challenges.
Regarding AI and machine learning innovations like ChatGPT and Bard, how do you view these technologies from a customer use case perspective? What new opportunities do they present?
I would like to offer two complementary perspectives. The first is that market shifts can significantly impact even the most established enterprises. In today's dynamic environment, no one is immune to change. Therefore, continuous innovation is not just an option but a necessity for survival. As for disruptive technologies like generative AI, it's crucial for a company like Kyndryl to deeply understand them. Recognising both the challenges and opportunities they present is the first step in ensuring these technologies don't disrupt our operations but instead are harnessed as opportunities.
Secondly, we are observing more opportunities than challenges with Generative AI. This technology is enabling us to address and solve customer problems that were previously hard to reach. For instance, without generative AI, addressing certain issues would require extensive research and development of skills in areas like large language models. Now, with generative AI models available as a service, we can focus on the engineering and problem-solving aspects to meet customer needs. This shift is creating numerous opportunities, which we are witnessing across various sectors. For example, in the tech industry, we are tackling real-world problems related to natural language processing by leveraging existing generative AI models.
Generative AI is democratising and levelling the playing field. It is narrowing the knowledge gap between different entities, making information and skills more accessible. However, effectively utilising generative AI does require expertise. For example, to fully exploit the capabilities of generative AI, a technical consultant needs a deep understanding of the technology and its applications. The true benefit derived from generative AI depends on one's own intelligence and domain knowledge.
We acknowledge the potential disruptions posed by generative AI, and are committed to understanding and utilising this technology to its fullest. This approach has opened up new opportunities, demonstrating the transformative power of generative AI in the tech landscape.
For large organisations with legacy technologies, balancing modernisation with existing systems is challenging. How do you advise CIOs or CISOs in managing this complexity, especially when considering newer technologies like Kubernetes and microservices?
We typically encounter three categories of customer needs. The first category consists of those who ‘lift and shift' their applications to the cloud. They benefit from the cloud's service model and an operational expenditure (OpEx) approach. The second category includes customers who develop cloud-native B2B (Business-to-Business) applications. This involves rewriting and refactoring their existing applications to fully leverage the cloud's capabilities. However, these applications usually witness predictable, and often low, loading.
The third category, which we discussed earlier, involves B2C (Business-to-Consumer) applications that experience significant load fluctuations. These applications require a high degree of cloud-native design, utilising the native features of hyperscaler cloud providers for financial optimisation. For instance, a media company might see a 100-fold fluctuation in user traffic between major sporting events. Such scenarios are impractical for on-premises infrastructure, which would require designing for peak usage.
Applications in this third category must be agile, maximising stateless code that can be efficiently run on cloud instances, such as spot instances or serverless functions that are cost-effective and require minimal management. The predictable parts of these applications require separate, algorithmically driven design considerations.
Our approach varies based on the customer's maturity, their specific use case, and the nature of their workload. For example, in sectors such as telecommunications and airlines, where we serve many leading companies across six continents, the demand is growing for modernising from traditional mainframe systems. Some clients opt to completely rewrite their systems for a cloud-native environment, while others prefer to emulate mainframe environments within the cloud, avoiding the cost and complexity of rewriting decades-old code.
Overall, our strategy is multidimensional, considering the customer's maturity level (the x-axis), their willingness to adopt new technology (the y-axis), and their budget or price point (the z-axis). This comprehensive approach allows us to tailor our solutions effectively to each customer's unique needs and circumstances.
In contexts like government organisations or OT-dependent sectors, how can they modernise, especially in air-gapped environments?
As an organisation that has diversified into next-generation networking with a focus on edge computing, we see a significant opportunity for modernising OT-dependent sectors, including in air-gapped environments. Edge computing plays a crucial role here, as it involves deploying computing power and connectivity close to the location where data is generated. This is particularly relevant for OT systems where real-time data processing and decision-making are critical.
For example, in air-gapped environments like those often found in government and critical infrastructure sectors, traditional cloud computing models aren't feasible due to security concerns. Here, edge computing allows for the processing of data on-site, reducing the need for external data transmission and enhancing security. This is crucial for sectors where data sensitivity and response times are paramount.
In practice, we have applied this approach in various sectors. For instance, for one of the largest chemical producers in the world, we implemented a private 4G network at the edge across four countries, significantly transforming their OT processes and business model. This helped in mitigating disruptions and enhancing performance in a secure, air-gapped environment.
Similarly, in India, we are working with a major telecom company to create a near-edge network across India for deploying latency-sensitive applications that need real-time data processing.
Also, in Europe, we have assisted a leading glass manufacturing company in enhancing their OT processes. We have implemented edge computing solutions near critical points such as blast furnaces to manage and visualise on-site intelligence. This bridges the gap between public cloud services and local operations, tailoring solutions to the specific needs of OT environments.
Our involvement in edge computing across these various sectors demonstrates our commitment to addressing the unique challenges of OT environments, including in air-gapped scenarios. We believe that by leveraging edge computing, we can offer innovative, secure, and efficient solutions tailored to the specific needs of these sectors.
How do you view India's progress in technology, considering the Digital India programme and indigenous technology efforts?
I believe we are certainly on the right track. Looking back from 2016 to 2023, India has seen a remarkable transformation in various sectors that has significantly impacted the lives of ordinary citizens. Take fintech, for example. In 2016, UPI was non-existent, but today, the number of UPI transactions stands at an impressive 11 billion a month, surpassing the combined total of the rest of the world.
India ranks third globally in fintech unicorns, a testament to the sector's tremendous innovation. At Kyndryl, we are contributing to this progress, tackling complex areas such as specialised lending. In India, lending instruments are varied and complex, especially in agriculture. Unlike home loans, where a bank assesses risk based on a borrower's income and expenses, agricultural lending involves analysing land ownership, crop types, yields, and even weather patterns from previous years to predict future risks.
In addition to fintech and microlending, we are actively involved in transforming industries such as education, aviation and steel manufacturing. Across many sectors in India, we see a wave of transformation, each with its own unique set of challenges and opportunities. This widespread change is indicative of India's growing importance in the global market, further evidenced by the fact that it is now a top innovation and R&D destination. The progress made thus far is undeniable and promising.