Harsh Vardhan is Global Head, Digital Innovation for Apollo Tyres, an international tire manufacturer and the leading tire brand in India, with global manufacturing units and global operations. It sells tires under its two global brands – Apollo and Bredestein – across 100+ countries through a network of branded, exclusive and multi-product outlets.
Vardhan is responsible for driving AI-first digital innovation, fostering both top-down and bottom-up innovation cultures, and for implementing other cutting-edge technologies such AR/VR, and digital twins to transform business processes, enhance manufacturing efficiency, and accelerate sustainability goals. Under his leadership, Apollo Tyres has launched six enterprise-grade Gen AI pilot projects: half of them with young companies based at Hyderabad’s T-Hub, one of the world’s largest innovation accelerators. Vardhan has opened a 2000 square foot mobility center of excellence (Digital Innovation Hub) for Apollo at the accelerator as part of an open innovation program focused on everything from supply chain optimization to manufacturing
Vardhan, who has a track record of collaborating with industry leaders and academia to advance open innovation and partner collaboration, was named one of India’s Economic Times’ Top 100 CIOs 2024 and e-GOV Top 40 CIOs 2024.
Prior to joining Apollo Tyres, Vardhan was a senior director at CapGemini, with the title of Chief Digital Transformation Leader, Data, AI, IoT and Multi-Cloud Modernization. Earlier in his career he served as Associate Product Director/Principal Architect for AI & IoT Multi-Cloud platform at Cognizant, an American multinational information technology services and consulting company.
Vardhan has a post graduate diploma in data science from India’s International Institute of Information Technology and a Bachelor of Engineering degree from the country’s National Institute of Technology. He recently spoke to The Innovator about open innovation and how to scale AI.
Q: How do you decide on where to apply Generative AI?
HV: We decided not to do AI or any tech adoption for the sake of adopting technology. It needs to be very clearly attached to a business function. We first identify a business case, then we try to create a return on investment (ROI) case. After that I speak to all our business unit heads across zones, regions and plants. With any breakthrough tech if there is a calculated ROI in one plant there is the potential for a bigger ROI across the globe. Cultural aspects need to be aligned to the business case. It is best to think big but start small in a place with the best fit. That said, it is important to note that no one technology will give me the business solution. My task is always about intelligent automation (IA) which is the fusion of artificial intelligence; niche digital technologies such as Generative AI, IIoT [Industrial Internet of Things], streaming analytics, etc.; continuous process improvement powered by TQM [total quality management], TPM [total productive maintenance] and Lean Six Sigma; and design thinking. When applied on top of clean data, process improvements and cross functional employee collaboration Gen AI is not just the icing on the cake, but an effective accelerator of Intelligent Automation.
Q: What can you tell us about your enterprise grade GenAI projects?
HV: One of the six significant enterprise grade Gen AI projects is focused on enabling our plant experts to take real time corrective measures to improve the efficiency and cycle time of high-speed fully automated machineries by supporting accelerated root cause diagnosis for the cause of deviations/loss of performance with extreme precision. This is a bold step towards Industry 5.0 and is completely in-house developed. This project is an integral part of our company’s very ambitious program to improve the capacity of top performing plants by sweating assets through AI/ML and digital transformation initiatives. This program is being anchored by a cross-functional team. The program is aimed at harnessing additional output in the range of 14%-18% in a plant where there is no low hanging fruit. Instead of the traditional approach of either setting up a greenfield operation or scaling up the operation through capex driven augmentation, the leadership has resorted to efficient asset sweating powered by digitalization/ AI/ ML / novel processes in alignment with the company’s larger sustainability goals. Our machines across plants are all IoT connected and are sending real time signals to the Cloud. This Gen AI-powered manufacturing reasoner platform is putting the capability in the hands of the operator to talk back to the machines and identify the focus area, which is pretty much like trying to find a needle in a haystack. There are thousands of parameters and AI can help us find which machine and what mechanics we need to adjust in real time. With this Gen AI platform, a shop floor operator doesn’t need to spend a lot of time on the IIoT dashboard to do the 5W1H analysis to find the bottleneck focus areas. They go and ask the question amongst the hundreds of machines: ‘Which one is having a problem, and what sort of problem is causing deviation/loss in efficiency?’ The machines give the answer. This happens at the operational and streaming level. Thanks to real stakeholders -i.e. plant experts and stakeholders- this advanced GenAI platform is continually getting fine-tuned and increasingly adopted in daily work.
We are also using GenAI and other technologies to accelerate the ambitious goal of flattening the HSE [Health, Safety and Environment] pyramid, so that over time we report zero potential serious injury or fatalities and zero serious lost time injuries. In another of our enterprise grade GenAI pilots we are accelerating the turnaround time for smart informed research and innovative product design for our R&D business unit.
Q: How is working with startups T-Hub helping you with intelligent automation?
HV: The leadership has taken a conscious decision to resort to driving Intelligent automation at scale and at speed through digital Innovation and transformation based on four key tenets, which are: open innovation, design thinking, sustainable innovation and AI-led innovation culture. To walk the talk of open innovation through startups, Apollo Tyres decided to open its Center of Digital Innovation hub (DIH) in T-hub Hyderabad. T-Hub is certainly helpful in assisting Apollo Tyres in driving open innovation, bottom-up innovation culture through design thinking programs, and in finding apt startups for different digital innovation projects.
Q: What have you learned in working with startups that might be helpful to other companies?
HV: Driving digital transformation and sustainable innovation through startups has its own unique advantages and disadvantages. Startups are typically specialized in a particular area. Their offerings are equipped with the latest tech, they have a nimble approach and are often an apt fit for a particular domain-focused pain point. They are typically ready to solve a problem with razor sharp focus in a relatively better turnaround time, but one must vet the potential startup from the lens of cultural fit, rigor and the scalability, security and configurability of their solutions. A lack of a structured funneling mechanism in partnering with startups can lead to innumerable issues.
Q: You are an advocate of opensource AI. What do you see as the benefits?
HV: In my individual capacity, I’m an A+ open-source contributor in the field of advanced AI/ML, X-Ops, and other niche digital tech. I endorse the industrial adoption of opensource AI at scale and at speed for the following good rationales:
- The widely known and revered open-source platforms host a vast collection of openly available models and datasets that anyone can access and utilize.
- An open-source approach fosters collaboration within the AI community, enabling researchers and developers to share their work, contribute improvements, and innovate collectively.
- The emphasis on open-source software and AI model allows organizations and individual developers to leverage advanced machine learning capabilities without needing substantial financial resources or proprietary technology.
- Open-source AI models help to mitigate the concentration of power among large tech companies that typically dominate the AI landscape.
- Open-source Gen AI models are increasingly focused on domain specific smaller models (SLMs) rather than LLMs —those that perform comparably to larger counterparts but require less computational power. This, in turn, ensures further enhanced accessibility. These smaller models can be run on personal devices without relying heavily on Cloud infrastructure, thus promoting privacy and reducing energy consumption associated with running large-scale AI systems.
One of the such widely known and revered open-source platform is Hugging Face , which was established in 2016 and has quickly evolved into a key player in the open-source AI community. Hugging Face is playing a crucial role in the democratization of AI through its commitment to open-source principles.
Q: What advice do you have for corporates that want to scale GenAI?
HV: Be picky about what problems you tackle. Think big, start small and scale fast with the right business objectives, cross functional teams and design thinking- led process improvements. Align cultural aspects with the business case. Ignore the buzz and hype around LLM models and don’t simply assume that costly LLMs will give you the better outcome. Use the right tech for the right problem.
This article is content that would normally only be available to subscribers. Sign up for a four-week free trial to see what you have been missing.
To access more of The Innovator’s Interview Of The Week articles click here.
