In a recent interaction with Express Computer, Rahul Kedia, Chief AI Officer, Sandvik Group IT, shares insights into the evolution of AI and its implementation across industries, particularly in India and Europe. He highlights a marked increase in curiosity about AI’s potential, particularly in manufacturing, spurred by the advent of tools like ChatGPT. Kedia discusses Sandvik’s strategic approach to integrating AI into their operations, focusing on both internal processes and external customer-facing applications.
Since you spend time in both Europe and India, how would you compare the adoption and integration of AI between the two regions?
I don’t see much difference between countries in AI adoption. Over the last two years, especially at Sandvik and other manufacturing companies in Sweden, there has been a significant increase in curiosity about how to use AI, especially with the emergence of tools like ChatGPT. People are eager to know how AI can be aligned with business outcomes, such as increasing revenue, reducing costs, mitigating risks, or improving customer experience.
At Sandvik, we’re a 165-year-old, decentralised manufacturing company, and couple of years back established a department focused on AI. We’re exploring how to wrap digital solutions around our top-tier physical products to create new revenue streams, especially in areas like mining equipment and metal cutting tools, where we are industry leaders.
Our approach to digitalisation is twofold: internal digitalisation for smarter processes and external digitalisation through products and services. For instance, AI is used for forecasting, pricing, visual inspections, and customer-facing applications like chatbots. We’re also seeing similar trends across the Swedish ecosystem.
Regarding India, from what I’ve observed, companies like ITC and Hindustan Unilever are doing a great job, particularly with centralised mandates that make driving change easier. In Sweden, we rely more on a “pull” strategy, where business units must see the value of AI before investing. It can feel slower, but it reflects cultural differences.
Given that Sandvik has a 165-year history and has experienced numerous technological advancements, how challenging was the process of change management, particularly transitioning from traditional technologies to AI?
Change management has been quite challenging, especially transitioning from traditional technologies to AI. For example, we have large distribution centres where one of the biggest cost drivers is the walking distance in the aisles. To address this, we developed a simulator to optimise the slotting and pickup process, helping reduce time and improve customer satisfaction. However, the real challenge is getting people, some of whom have been with the company for 30-40 years, to adopt new methods. They often feel they know how to do things best, and question how AI can improve the process.
It’s not just about storing items; it’s also about understanding product affinities, like storing an iPhone with its case and accessories. Change management is never easy, even with advancements like the generative AI tool we created for our 3,000 service technicians. While the technology is ready, getting people to use it is another hurdle.
The key is working with business leaders. If they see a significant return on investment (ROI) from an AI initiative, like reducing costs or improving operational efficiency, they will take ownership of the change management process. It’s important to focus on use cases that have both technical feasibility and a strong business impact.
Do you think that the incorporation of these technologies like AI and automation could lead to potential job losses, even if only temporarily?
To be honest, it’s already happening. Job losses due to AI and automation are not a future concern—they’re a reality now, just not at Sandvik yet. The shift is ongoing, and it’s about how you embrace these technologies. For instance, AI won’t replace jobs, but people using AI will replace those who aren’t. That’s the point I emphasize to employees: embrace AI to work smarter, faster, and more efficiently. You need to upskill constantly.
For example, we had 800 applicants for a managing director role in India, and our recruitment team used an AI-powered candidate screening tool to filter the top 50 candidates. This is how AI can make processes more efficient. New roles, like Chief AI Officer, didn’t exist a few years ago. Now, it’s about showing business impact.
How do you see the ongoing AI hype sustaining itself, and what role does data play in ensuring AI delivers real business outcomes despite challenges with data quality?
Typically, we refer to the Gartner Hype Cycle, which shows the curve of technology trends. Most technologies hit the peak of inflated expectations and then decline. However, with AI, over the past two years, it has remained at that peak. Every couple of months, something new emerges—first text, then images, audio, video, and now multi-modal AI. That constant innovation is why AI remains a hot topic.
In fact, even local shop in India, people are likely talking about ChatGPT. The hype is everywhere, but I can’t predict exactly what will happen in the next 4-5 years. That said, it’s a great opportunity for AI leaders to turn that hype into real business outcomes.
One key point is that AI without data is useless. Many companies focus on AI-ready data rather than perfect data because in reality, you won’t always have flawless data. You’ll have a mix—some good, some bad. So, instead of spending the next five years solely improving data quality, I advise working on AI initiatives while improving your data in parallel.
I believe you would accept that another challenge is the quality of data that are being fed to the models and their potential biases.
Yes, it reflects human biases. We’ve been biased for decades. For example, when I talked to someone about training ChatGPT, I pointed out that if you ask it for an image of a shop or a road, it often gives a westernised version, not an Asian or South Asian one. This is because of the data it has been fed. It’s all about data. If you ask for a nurse, it will likely show a woman, or for an engineer, a man. At airports in the U.S., people of colour, particularly from Africa, were often stopped and checked. Now, AI is mirroring those same biases. It’s a chance for us to improve our data and our thinking.
Is there anything in particular that excites you and that you’re eager to take charge of in the next few months or years?
I’m excited to work more with digital products and services. That’s where my passion lies and where I spend most of my time, as it allows us to impact the top line. We are experiencing saturation in certain areas of our business, particularly with tooling, due to electrification. As we require fewer products, we need to disrupt the market; if we don’t, someone else will. We’re currently pursuing several digital initiatives, which I believe have the potential to be game-changing, particularly with AI.