February, 2025

9 mins read

The Future of AI is Lean with a Shift from LLMs to SLMs

Pradipto Mukherjee - Senior GM Finance - Global Manufacturing Operations, Glenmark Pharmaceuticals 09 Mar 2025

As AI advances at breakneck speed, the future may belong not to bigger models, but to smaller, smarter ones. Pradipto Mukherjee, Senior GM Finance - Global Manufacturing Operations at Glenmark Pharmaceuticals, predicts a shift from large language models (LLMs) to small language models (SLMs) for greater efficiency and scalability. In conversation with Sai Moukthik Konduru, he unpacks the challenges of post-merger integration, the cultural nuances of global consulting, and why prioritising learning over quick gains can shape a stronger career.

The Future of AI is Lean with a Shift from LLMs to SLMs

What drew you to consulting as a profession? If it wasn’t consulting, what other domain would you have preferred to work?

What appealed to me about consulting is the ability to be analytical and solve problems across industries and functions. My focus has always been on solving business problems using technology. Problem-solving and exposure across different industries drew me to consulting. If not consulting, I’d probably be in finance because that field is also somewhat analytical. Definitely not sales, and definitely not insurance.

Any particular reason for that?

Sales essentially requires you to be likable to someone, right? You need to have the ability to cater to their point of view or their taste. I’ve always had a very definitive point of view regarding problems, and I didn’t always feel comfortable needing to cater to other stakeholders. That’s just a personal trait of mine.

You’ve worked with both Indian and foreign multinationals. What key differences have you noticed in their approach to consulting, organisational culture and so on?

I don’t think there’s a major difference in their approach to consulting from a methodology standpoint. However, let’s take the example of firms like IBM and Accenture versus Infosys and Cognizant. IBM and Accenture have been in consulting for a much longer duration than the consulting arms of Indian firms. So, what they bring to the table is depth of expertise due to their years in the industry; their subject matter experts tend to have deeper domain knowledge and more extensive experience.

In Indian companies, we tend to be a more close-knit group, and access to clients is often much higher for Indian consulting service providers. In global firms like IBM and Accenture, the global delivery model means you’re generally supporting a region or geography, whereas in India, it’s much more accessible. So, in summary, the difference is that global firms offer more depth and longer industry expertise, allowing you to learn quickly and go deeper in consulting. In Indian consulting firms, you have greater access to clients and opportunities. From a methodological or consulting approach, there’s no major difference.

You’ve worked across various domains in your career, from banking to technology. What skills or strategies have helped you make this transition smoothly and be successful throughout?

Essentially, in consulting, there’s an expectation to grow deep in a specific domain for the client you’re working on. That takes personal effort to understand the domain and gain expertise. When you’re consulting in a company for, say, three months, the client has its own domain experts who understand the landscape better than you can. You can always ask deep domain questions from the client and upskill yourself. I’ve also been fortunate to have colleagues with deep domain expertise, so I learned a lot from them. So, it’s really a combination of self-learning, learning from clients, and learning from colleagues.

While mergers are often celebrated and make big headlines, there’s very little insight on what happens post-merger and how companies make the merger successful. Having worked extensively in this area, could you share your insights?

Absolutely. Post-merger integration, especially with technology, is an often-overlooked aspect. Everyone celebrates the merger but doesn’t always appreciate the immense effort needed for successful integration. In fact, research shows that less than a third of companies actually manage to profit from mergers, and the reasons for that are widespread.

First, with technology integration, what happens is that the CIO of the acquiring company is often not at the decision-making table when it comes to setting cost-saving or synergy targets. So, they’re handed synergy targets without any input, which can be a huge oversight. They really should have a seat at the table when defining these targets.

Second, there’s often inadequate due diligence, especially regarding IT systems. Many companies jump into execution without proper planning and due diligence. A detailed and dedicated pre- and post-merger integration team with a well-thought-out roadmap is essential.

And third, there’s culture integration. Merging two different working cultures, even within the same geography, is no small feat. Change management and culture harmonisation are essential. Ensuring the adoption of new ways of working and a harmonised culture is incredibly important.

To summarise, not involving IT in decision-making, lack of proper planning and due diligence, and inadequate focus on culture integration are the three critical issues in post-merger integration. It’s great that more people are paying attention to this now, but it’s still underappreciated.

Over your career, you must have worked on multiple projects. Is there one particularly interesting project that stands out? Feel free to share more than one.

I’d actually like to talk about a project where I didn’t succeed, but I learned a lot. I was the first person from IBM India to go to the Nordics on a project — in Sweden, specifically. The client was a major telecom company, and I was tasked with organisational change management and cultural transformation.

What I didn’t realise initially — but understood in hindsight — was that decision-making in the Nordics is very different from that in the US or India. In the US, decision-making is more top-down, and in India, it’s similar in many cases. However, in the Nordics and other European countries, it’s much more consensus-driven.

When I conducted my change management and cultural transformation work there, I primarily focused on the chief executive and didn’t focus enough on his direct reports. Later, I realised that his direct reports actually had a much greater say in decision-making than I anticipated, which ultimately led the company to end my engagement sooner than expected. This taught me the importance of understanding the cultural nuances of an organisation and its decision-making process. It’s not just hard skills that make a consultant successful; it’s also soft skills.

This experience happened early in my career, and it gave me a better perspective for future projects. So, while it wasn’t a success from a results perspective, it was an invaluable learning experience. Sometimes we win; sometimes we learn.

With rapid advancements in technology, particularly in the use of data across various applications and recent breakthroughs in generative AI, how do you see these changes shaping industries and the global economy over the next decade?

That’s a great question. I see several key trends impacting industries and the global economy, particularly in AI and generative AI, which, as you said, is the ‘flavour of the town’ right now. In terms of generative AI, it’s fascinating yet challenging. On one end of the spectrum, there’s a lot of FOMO (fear of missing out) among companies. On the other end, many are getting stuck in a cycle of proof of concepts (POCs) that fail to make it to production, largely due to the high cost associated with deploying generative AI models, which is still quite prohibitive compared to traditional AI or data projects.

The second major area to focus on is data management and governance. Ensuring the quality, availability, and accessibility of data is crucial for both AI and generative AI models. These models require vast amounts of data, so improving data governance and creating robust frameworks to prevent issues like AI hallucination is critical. We’ll likely see a shift from large language models (LLMs) to small language models (SLMs) to make these technologies more cost-effective and scalable. Additionally, the data governance and management side will need more attention to prepare training data for LLMs and beyond.

The next big trend is around customer experience. Customers are increasingly expecting consistent, high-quality experiences across all channels. The same convenience and seamless service they receive on platforms like Amazon will be expected across in-person experiences as well, from retail to grocery shopping. This demand for a consistent, exceptional customer experience will drive companies to innovate their digital and physical customer interactions.

Lastly, quantum computing is another field to watch. It’s still largely experimental, but it’s likely to become more mainstream in the coming years. Its potential to revolutionise computing power could transform industries and applications in ways we’re only beginning to understand.

So, to summarise, I see three major areas: improvements in data governance and AI, the push for a unified customer experience across platforms, and the rise of quantum computing. These will be some of the primary forces shaping technology and its impact on the economy.

You are an IIMB alum — a dream school for many. Reflecting on your MBA journey, are there any specific lessons or experiences from IIMB that continue to influence you today?

I think one thing is the ability to scale up quickly, a skill you’ll appreciate, especially with the compressed nature of your one-year programmes. At IIMB, I learned the importance of rapid learning, particularly with the three-month semester structure. While we couldn’t master every subject in-depth, learning how to grasp concepts quickly became essential. Some of my classmates could do this even faster than I could, which was inspiring. Another thing was applying concepts practically. In engineering, we didn’t have many assignments or practical demonstrations. But at IIMB, presenting recommendations to a critical audience, even if they were classmates, was part of the experience.

Additionally, class participation was something daunting for me as I’m a bit soft-spoken. But that pressure to participate — both during and after class — helped me build confidence in voicing my thoughts.

Let’s wrap up with a final question. What advice would you give to your 25-year-old self, fresh out of IIM Bangalore?

The main advice would be that learning often outweighs money. With a student loan, it’s tough to prioritise learning, but it’s a balance. I had three options after graduation, and I chose the firm that was the most lucrative in terms of remuneration. But in hindsight, two of the other options may have offered better learning opportunities. It’s a tough choice, but I’m sure you all will make the right decisions.