I'm the Technology Editor at Forbes India and I love writing about all things tech. Explaining the big picture, where tech meets business and society, is what drives me. I don't get to do that every day, but I live for those well-crafted stories, written simply, sans jargon.
Satish HC, Executive vice president and co-head, delivery, Infosys
Satish HC, executive vice president and co-head, delivery, at Infosys, appeared on The Daily Tech Conversation, recently. He spoke about how Infosys developed Topaz, the experience of early adopters, and how the Bengaluru IT company is using these technologies in-house to raise the bar on its own performance. Edited excerpts:
Q. Tell us about Infosys Topaz—what it is, the idea behind it, and how it works with your cloud suite Cobalt. Infosys Topaz is an AI initiative that also incorporates generative AI. Over the past five to six years, we have gained extensive experience in delivering AI solutions, transitioning from an AI strategy to a use-case-based approach, and now focusing on scaling AI within enterprises.
With generative AI entering the picture, we believe it presents a significant opportunity for AI evolution. We carefully choose our brands at Infosys, and Topaz represents our expertise and business potential in this domain. Topaz encompasses a range of AI services, solutions, platforms, and partnerships, consolidating our offerings under a single brand to address the vast opportunities ahead.
The virtuous cycle of digitisation and data creation drives the need for a strong digital and cloud foundation to leverage the power of AI. While some clients may have already established these foundations, many are still working on them or planning to invest in AI.
In those cases, we provide both Cobalt (Infosys’s cloud suite) and Topaz solutions, supporting their digital and cloud journeys while simultaneously building their AI capabilities. The interplay between Cobalt and Topaz depends on the client's investment and use cases, but overall, both technologies will work together to drive digitization and AI application. It's a continuous process of reinforcement and advancement.
Q. Give us some specifics in terms of the main features and capabilities. Let me begin by talking about the frameworks we have developed based on our experiences. One such framework is the digital brain, which we created to harness real-time signals and drive value for stakeholders. With the advent of generative AI, we are evolving this framework further to amplify its capabilities.
We also took an early pivot towards a poly-cloud approach, believing in the importance of leveraging different stacks for different areas of AI expertise. This led us to create an AI cloud and provide prescriptive advice on the best digital railroad for clients to enhance their AI capabilities. We developed ML Ops (machine learning operations) and AI Ops platform capabilities to work at digital clock speed and enable fast-paced AI development. Autonomous data engineering became another crucial aspect, ensuring the ingestion of high-quality data at high speed.
Alongside these frameworks, we have built industry intelligence clouds and specific functional capabilities like digital claims assistant, document distinction, distilled twins for product engineering, and smart warehousing. We possess deep expertise in generative AI, particularly in vision, speech, conversation, text, and transaction AI.
Beyond traditional RPA and automation, cognitive automation has become the focus of our discussions. Overall, our extensive frameworks, solutions, and expertise in generative AI make us well-equipped to cater to the evolving AI landscape.
Q. Can you talk about two examples of early adopters. One of our notable projects involved a British bank where we added value in two ways. First, we used an AI engine to monitor customer support interactions in real-time and guide the agents. This proactive approach ensured better customer experiences by identifying if the conversation was going in the wrong direction and suggesting ways to improve it. Additionally, we analysed transactions retrospectively to continuously enhance service efficiency.
Another powerful example was our ability to predict why a customer was calling the bank based on their transaction history. With a 360-degree view of their interactions, we automatically routed the call to the appropriate agent, reducing waiting times and improving customer experiences. These were the examples mentioned in the press release.
For a telecom company, we developed an end-to-end framework to aid visually and hearing-impaired customers. This solution provided visual captioning and generated audio descriptions of visual scenes in near real-time for visually-impaired customers. Similarly, for the hearing-impaired, we generated automated closed captioning for live streaming content. These cognitive enhancements helped serve an underserved community and provided them with near real-time accessibility.
In another case, we worked with a multinational food company to improve customer engagement. By creating a customer preference graph, we engaged consumers with recipes and personalised recommendations based on their preferences. This resulted in a 30 percent increase in consumer engagement, a 23 percent decrease in bounce rate, and an 80 percent increase in personalisation.
Further, we leverage AI for social media analytics in sports retail, managing brand sentiment and engagement with 30 percent less effort. These examples demonstrate how we are applying AI to drive value and enhance efficiency in various business scenarios, both for our clients and our own services. By harnessing generative AI, we have become more productive and empowered to deliver faster results.
Q. Tell us about how you’ve addressed concerns around the pitfalls and dangers of AI, and generative AI, in building Topaz. In the journey of adopting new technologies like AI, there is a need for experimentation and exploration to establish their value and potential. Regulation plays a crucial role in addressing pitfalls and ensuring responsible usage, as seen in domains like clinical trials and financial services.
Similarly, social media has demonstrated its power to empower individuals but also raised concerns regarding harmony and trust. AI, like any technology, has its own challenges, such as the need for high-quality and certified data, explainability of decisions, and human oversight.
We focus on building AI solutions with emphasis on data quality, explainability, and human involvement. We acknowledge the importance of legal and ethical issues, such as intellectual property rights and content generation attribution. While there are still gaps and ongoing progress in AI regulations, we believe in working within the existing frameworks and guiding clients to operate within ethical boundaries.
The aim is not to ban AI but to understand and evolve the regulations that will harness its potential while addressing concerns. By considering factors like explainability, auditability, human involvement, and data foundations, we strive to build AI solutions that align with organizational values and can withstand scrutiny.
Q. How are you using these AI tools within Infosys? Apart from client projects, we also apply these technologies to our own services. For example, we use generative AI to extract knowledge hidden in documents and accelerate learning during knowledge transfer transitions. We are also using generative AI to convert Cobol code to Java, optimising reengineering and modernisation projects for clients.
Nandan Nilekani (chairman of Infosys) initiated our digital transformation in 2017, making us a digital-native enterprise. Now we are embracing an AI-first approach within our organisation. We have been working on implementing AI for about three months, with the goal of establishing a core AI infrastructure within six months.
One way we are using AI is by creating AI twins for individuals undergoing training. For instance, when writing code, an AI engine provides real-time feedback on code quality and helps accelerate learning. Learning has always been a significant investment for us, and now we are applying AI to enhance the learning experience.
We are incorporating AI in various aspects of our software engineering processes. Examples include knowledge transfer, generative AI for code generation, and cognitive automation to handle complex and evolving scenarios. We believe that AI will boost productivity and become integral to software engineering and enterprise architecture. To facilitate this, we are investing in building a ‘digital brain’ that can orchestrate and respond to real-time signals based on learning.
With the advent of generative AI, our software developers will code faster and gain access to qualified answers and metrics. They can also evolve into consultants, leveraging AI for research and analysis. The advancement of technology will empower individuals to pursue higher-order tasks and tap into their potential.
We anticipate a future where business users can write specifications, and AI will generate the corresponding code, simplifying system orchestration. This will lead to increased autonomy and automation, while complex scenarios will still require the expertise of full-stack programmers.
This evolution will gradually reshape job roles and tasks, allowing our technology professionals to focus on higher-value activities. Business users will have greater autonomy in leveraging technology, reducing dependency on technical intermediaries. The potential for innovation and productivity gains is immense. For instance, creating presentations can be simplified by generating them using AI tools, liberating individuals and enabling them to deliver beyond their previous capabilities.
This transformation will have a long-term impact on jobs and learning. Formal institutions may become less essential for basic education as learning methods evolve. We are entering an era where right-brained individuals who can reimagine the world and contribute to a better society will thrive. This journey will span several decades, driven by technology as an enabler.