We recently had an in-depth discussion with Sreekanth Menon, a seasoned industry expert who heads the AI/ML practice at Genpact Analytics. Sreekanth is at the helm of global AI/ML projects delivery and has spent over two decades nurturing innovation and industry expertise. He has been instrumental in incubating and launching more than 50 advanced analytics solutions in the global market and has driven business transformation in partnership with Fortune 500 clients through innovative AI-led solutions and practices.

In this exclusive interview, Sreekanth will share invaluable insights into Genpact’s AI strategy, which is founded on four pillars: AI at Scale, AI-Driven Insights, AI-Enabled Autonomous Process, and AI for Operations. He will also shed light on Genpact’s generative AI strategy, the various use cases they are developing, and their vision for the future of Large Language Models (LLMs).

We will explore the partnerships Genpact is fostering to operationalize AI more efficiently and understand how they are equipping their talent pool to adapt to the rapid changes in the AI landscape.

What is your AI strategy in Genpact?

At Genpact, over the years we have developed and refined our AI capabilities, enabling us to create innovative, industry-specific solutions for our clients. Engagement models across industries are changing rapidly and more so with the advent of generative AI. Today, enterprises would like to offer their services – everything, everywhere, and all the time. While they still desire an even mix of traditional and modern tools. Genpact’s AI strategy has resulted in state-of-the-art frameworks accommodating such hybrid demands.

Genpact’s AI strategy is founded on 4 pillars:

• AI at Scale: Building full-stack AI applications and multi-year AI/ML managed services with DataOps and MLOps
• AI-Driven Insights: Descriptive insights to predictive insights to enable expansion in prioritized accounts.
• AI-Enabled Autonomous Process: Automate, Optimize and Redefine business processes, operating, and customer models through an AI lens.
• AI for Operations: Infusion of AI for Ops transformation and modernization.

These foundations are reinforced with agile and ethical frameworks that allow AI developers to bridge the gap between ideas and implementations.

“Nurturing Responsible AI adoption is a key strategic imperative at Genpact.”

Do you have a GenAI strategy in place and how are you looking at leveraging GenAI for your clients?

Our gen AI strategy relies on a paradigm of democratization through incubation with a strong footing in responsible AI. Genpact’s generative AI strategy considers different stages of adoption: incubation of processes, people, tools and technology, and eventually, democratization. This allows us to establish large language model (LLM) centers of excellence (CoEs), identify the technology foundation, enhance the tool stack, streamline processes, and empower the workforce with self-serve gen AI apps across the enterprise.

Centers of excellence can be leveraged as a change management hub to design, integrate, scale, and democratize prototypes into enterprise-grade solutions. They help incubate in-house employees into the roles of prompt engineers, prompt compliant checkers, customer protection officers, and other such relevant roles. These centers serve as a hub to design, integrate and scale the generative AI prototypes to enterprise-grade solutions.

Genpact’s gen AI offerings for our clients follow a three-pronged approach:

1. Generative AI for Enterprise LLMs for building fine-tuned custom foundation models with DataOps and FMOps to enable the adoption of generative AI applications in enterprises.
2. Generative AI enabled business processes for automating, optimizing, and redefining business processes with generative AI capabilities of search, generate, classify, cluster, summarize and extract.
3. Generative AI for Operations for augmenting workflows for developers and analysts in technology services and digital operations.

What kind of use cases are you developing today in Genpact?

Genpact’s gen AI competencies flourish at the intersection of various industries and their service lines. If we take the healthcare sector, for example, then we are looking at hyper-personalized care and patient experience, which revolves around value-based care. Gen AI’s potential for banking can’t be overstated. LLM-powered apps can help banks grow revenue, manage risk, enhance customer experience, drive innovation, and lower costs.

The same goes for insurance companies, which are very document-driven companies. Underwriters spend 30–40% of their time on administrative jobs, according to one estimate. The exhaustive nature of these tasks can be alleviated with LLM applications.

These offerings allow enterprises to leverage the power of LLMs with great ease. This can be mainly attributed to our strategy, which is carefully weaved within a tried and tested Responsible AI framework of Genpact. These frameworks allow our customers to establish ethics-as-a-service in the context of gen AI and prevent risks.

What is your view on the future of LLMs?

Traditional AI/ML solutions at the enterprise level is staring at a major overhaul. Language is the interface. Currently, enterprises are trying to wrap their heads around prompt engineering. But the complexity of prompt engineering varies. At Genpact, the approach towards enterprise-grade gen AI solutions traverses through three stages:

• prompt tuning,
• few shot tuning, and
• fine-tuning the LLMs.

Moreover, as foundation models mature, more emphasis will be placed on protecting sensitive information. The strategies would be centered around ensuring that the company has the right people and management in place to stay competitive and maximize its AI investments. The future of LLMs would more or less trace the following journey:

• Multi-modal mastery: Improved natural language understanding, problem-solving capabilities, and integration of various data types.
• Dynamic Knowledge Integration: Enhanced learning with fewer examples for reduced training time and resources, along with real-time adaptation to new information and seamless integration with existing knowledge.
• Hyper-Personalization: Improved responsiveness and adaptability to user inputs, changing requirements, and customized solutions.
• Responsible AI integration: Ensuring safety, ethics, and collaboration in AI development.

What partnerships are you driving as part of AI @ Genpact?

Genpact has had strategic alliances with pioneers and leaders of the industry throughout its journey. We have recently joined forces with two of the biggest players in the AI industry. For instance, in partnership with AWS, Genpact is helping firms across the globe accelerate their digital transformation. Whereas an alliance with Dataiku is helping us build a differentiated solution that addresses major challenges faced by organizations in the implementation of MLOps and responsible AI at scale, such as data governance, model management, and compliance requirements.

Genpact’s partnership is aimed at enabling organizations to operationalize AI more efficiently. In the gen AI space, Genpact is exploring opportunities with the hyper scalers on co-innovation.

How are you equipping your talent to drive this sudden change in the ecosystem?

Upskilling is Genpact’s forte. We have been leading the upskilling charge for over a decade now. Our initiatives cover all things AI/ML. The reason behind our success can be attributed to our learning platform, Genome and our ML Incubator.

Genpact’s Genome is a learning platform that flaunts a curated list of coursework tailored to keep our over 100,000-strong workforce up to date with the trends of global digital dividends. Genome’s gen AI channel, for example, is designed to help employees understand the rubrics of how LLMs work and how they can be applied to build enterprise-grade solutions. Genpact’s ML Incubator is a meticulously structured  flagship hybrid learning program that drives contextualized learning for data science aspirants alongside their regular workflow. The program includes data science, data engineering, and augmented intelligence Advisory, each one of them a critical skill in today’s data/analytics landscape. The program caters to the scale at which a reskilling/upskilling effort is required for a large global digital organization to be able to have a ready pool of resources with AI/ML skills.