AI-Led Transformation in Automotive Manufacturing

Key highlights

What this conversation covers

Can you share your professional journey and your current responsibilities as the Group CIO of Krishna Group?

I completed my post-graduation in computer applications. I hold many certifications in the IT domain to keep updated on technology trends, and I am also a certified IT security auditor.

I have diversified and rich experience of more than 30 years in driving initiatives in IT strategy, implementing most of the renowned ERPs, IT security, business continuity planning and IT infrastructure designs, digital transformation, and Industry 4.0 implementation with automation of the shop floor in the manufacturing, power, infrastructure and shipping domains.

I have received several prestigious recognitions at the IDC Insights Awards, IDG CIO 100 Award, CIO SAMMAN by CIO & Leader, BW CIO Award, ET VIO Award, Magnificent CIOs of India Award, CIO Power List recognition, Cyber Smart CIO Award, Cloud Champs award by CIO Axis and many more.

I have been a panellist in various forums including manufacturing summits, ET CIO, 9dot9 Media, Bitstream Media, Polycom, CXOTV, Dun & Bradstreet, UBS Forum, IBM Think, and Fuelling the Future by Microsoft and Redington, to name some.

With Krishna Group I lead organisation-wide technology initiatives and drive Industry 4.0 adoption through automation of processes: use of AI through AR/VR projects, AI adoption for product inspection, integrating the machines for real-time monitoring and predictive maintenance initiatives, and creating data visualisation for better decision-making for top management with data enabled decisions. We built an in-house MES (Manufacturing Execution System) for sequence manufacturing and JIT finished goods creation with full traceability from start to end, plus full proofing of data security for the organisation-wide data assets, along with keeping the lights on for routine IT applications and uptime of the LAN and WAN setup using SD-WAN and cloud adoption.

How has your leadership role evolved with the growing influence of technology in the manufacturing industry?

The CIO's role has evolved from keeping the lights on to being an IT enabler and a strategic partner with the business, driving business value through the adoption of industry-specific right technology to maximise company-wide success and bring business efficiency. Real-time process monitoring for faster and data driven decision making, building an agile organisation embracing automation and business transformation, keeping in mind the internal and external customer's needs.

How is AI currently being leveraged in your organisation's manufacturing operations, such as improving supply chain efficiency, quality control, or predictive maintenance?

As AI is evolving day by day, we at Krishna Group are also adopting it for quality control using camera-based computer vision, which can detect minor abnormalities that might otherwise go unnoticed, thereby reducing waste and ensuring high-quality output in the manufacturing process. We use AI for predictive maintenance with IoT for monitoring machine behaviour and parameters, and AR/VR for manpower training on various aspects. These are some of the initial initiatives undertaken, and with fruitful results. As the applications stabilise we will horizontally deploy in other plants with further enhancements and faster deployments.

In your view, what are the most impactful areas within the automotive manufacturing industry where AI adoption can drive significant transformation?

AI is driving innovation at every turn in the automotive manufacturing industry. It is different for OEMs and different for component manufacturers, but AI is transforming how vehicles are produced, right from operations to service, predictive maintenance, autonomous driving, ADAS, supply chain automation, fleet management and customer experience. Automotive companies are leveraging AI to enhance efficiency, safety, and customer experience using AI/ML, AR/VR and digital twins. Component manufacturers can use it for increasing operational efficiencies through automated training. Computer and machine vision systems can automatically inspect and identify defects or anomalies in products during the manufacturing process, ensuring consistent quality control and leading to waste reduction. Real-time monitoring improves the machine's availability to sweat the assets and bring down the capex.

How do you ensure the availability and accuracy of data to enable AI-driven decision-making in manufacturing processes?

AI-driven decision-making in manufacturing processes involves a lot of data cleansing, preprocessing, and augmentation. Once data accuracy is achieved, robust evaluation metrics need to be in place to assess the performance of AI models.

With more and more digital dependency, there is a need to maintain high data quality by protecting it from unauthorised access and potential corruption. The concept of CIA (confidentiality, integrity and availability) has to be on top as the dependency on the data increases and any disruption can bring down the whole operations.

Hence, we are building a robust data governance framework to define data quality standards, processes, and roles. Zero trust data security implementation with multi-factor authentication are some technologies we are exploring for further fortification.

What role does data integration across departments (production, supply chain, and customer service) play in supporting AI and analytics initiatives?

Data integration is crucial and essential because it consolidates information from siloed sources into a unified, coherent and accurate view. It effectively solves the challenge of handling data from multiple providers, it breaks down data silos, ensuring that all parts of the organisation have access to consistent and accurate data. That enables overall data quality for a smooth flow of information and improved analytical capabilities that foster better decision-making. AI can only give the right insights or make informed decisions based on authentic data.

What metrics or KPIs do you prioritise to measure the effectiveness of AI-driven initiatives, such as improvements in efficiency, cost reduction, or quality enhancements?

With quantification of tangible and intangible benefits of AI projects with real-time monitoring, we measure the efficiencies, optimisation of machine utilisation (through machine running, idle time, breakdowns, tool change time monitoring), energy consumption, traceability and defect analysis. We get quality enhancements which lead to cost reduction and better customer experience. With AI-driven training and machine maintenance, turnaround times can be improved, and AI can contribute to safety and ESG implementations going forward.

How do you evaluate the ROI of implementing AI technologies in operations and supply chain management?

First and foremost is adoption rate, followed by employee experience and ease of use, performance of the AI model, and how many false positives. If it is a customer-facing application, then the customer's experience: quality improvements in the product delivered, improvement in time of delivery on designs, product improvements. Better monitoring and planning of material flow in the whole value chain leads to JIT inventory, visibility of material in transit, and quality of material received, with integration with a partner for inline production and quality monitoring.

What have been the key challenges in adopting AI technologies in a traditional manufacturing environment, such as talent readiness, legacy systems, or cultural resistance?

One of the primary challenges is the high initial cost associated with AI technology, making investment in hardware, software and infrastructure significantly critical.

There are limited use cases in manufacturing, a lack of awareness of manufacturing applications for the technology, and long ROI and results for new implementations.

Much legacy equipment exists in manufacturing settings, and integrating new technologies with a lack of interoperability between different systems is costly and time-consuming.

Over a period of time, unplanned downtime can hamper production, as reliance on AI can bring complacency, with AI lacking the ability to think of outside-the-box solutions, which is the USP of human intelligence in difficult situations.

Security needs and limited bandwidth from ISPs at remote manufacturing locations hinder the move to cloud adoption and force investment in edge computing hybrid infrastructure.

How do you manage the balance between automation and maintaining a skilled human workforce in a technology-driven manufacturing environment?

With limited use cases available in automation in a manufacturing environment, along with a crunch of skilled manpower, we take help from external consultants for solution design with continuous involvement of our teams, get the solution implemented and train teams throughout the process for operations and model teaching, fostering a learning culture which ensures that the workforce remains agile and ready to embrace evolving technologies.

What emerging AI trends or technologies do you believe will have the most transformative impact on automotive manufacturing in the next 3-5 years?

Autonomous driving, prescriptive maintenance, and sustainable practices are becoming a reality, with AI transforming the manufacturing process by automating manufacturing through robotics and enabling real-time data collection and analysis, leading to faster decision-making, streamlining production and enhancing quality control.

AI reduces the time required for design approval and sanction by enabling a quicker and better design workflow. Additionally, AI aids manufacturers in producing a variety of designs for improved product concepts and processes for autonomous vehicles, simulations on crash tests, product fitments and looks.

With more and more advancements and adoption, new ways of AI implementation will emerge, changing the views we hold today.

How do you see technologies like digital twins, machine vision, or generative design influencing the way manufacturing processes are optimised?

By embracing digital twins along with AI, IIoT and VR, one can gain deep insights into operations, improve decision-making, and enhance productivity through the creation of a virtual sandbox of the plant, instead of investing in assets for trials having multiple iterations.

Computer and machine vision systems are a technology that enables machines to perceive, analyse, and interpret visual information from the surrounding environment.

It involves the use of cameras or other imaging devices to capture images or videos and then applying advanced algorithms and techniques to extract meaningful information from these visual inputs, having wide implementation in quality inspection in manufacturing, robotics and automation, object recognition, safety, medical imaging and prescriptive remedy, and in agriculture, gaming and training.

Generative design is an innovative approach that involves using algorithms and artificial intelligence to generate multiple design options based on specified parameters and constraints.

Designers can explore numerous design alternatives quickly and efficiently. Generative design tools typically generate a multitude of design options, which can then be evaluated based on specific criteria. Complex geometric shapes and structures that may be difficult to conceive manually can be done easily. These designs can leverage the full potential of advanced manufacturing techniques such as additive manufacturing (3D printing).

Lightweight products and components lead to material savings and competitive advantage through accelerated innovation, and bring a product to market faster through improved efficiency by enabling multiple processes to take place simultaneously instead of traditional sequential processes.

It will hasten the use of additive manufacturing by creating complex designs for pilot projects, until the cost of additive manufacturing is reduced enough to have full small lot production.

How do you ensure responsible AI adoption in your organisation, particularly in terms of sustainability, data privacy, and fair practices, while scaling AI in manufacturing?

The policy is being drafted as of now, but clear communication to developers, testers, their managers, and project owners has been given to adopt best practices during development and implementation. Currently, we are using AI for machine data analysis where biases cannot generally happen, which is a big fear. For human-based decision-making through AI, if biased data inputs are given to AI and ML then results can be an issue. A basic level of transparency must exist with respect to development that affects decisions on humans, like medical records, autonomous driving, insurance claims and loan approvals, hence a third-party audit should be done for the AI itself and how it behaves. That means documenting all training and test data sets, the processes used to train a machine learning AI and the algorithms used, and ensuring explainable AI that is sufficiently understandable to humans so that the AI's decisions and impacts are accepted without question. On sustainability, AI's predictive maintenance capabilities can help avoid major issues by spotting problems early, optimising energy consumption and extra material consumption, and controlling rejections through early warning on product quality deterioration.

Lastly, what insights or recommendations would you like to see highlighted to guide businesses and leaders in leveraging AI effectively?

All my fellow CIOs are capable enough to make their own decisions, but I can only say that they should start with the problem to be solved, not the technology to be applied.

Choose with a focus on existing and mature AI technologies, not speculative promises or external pressures. Go for a proof of concept, start small and ramp up, fine-tuning the models to have better results. All the best to all friends in the endeavour to adopt AI.

AI systems need a variety of high-quality data sources. It can be difficult to ensure that data adequately depicts the different situations that cars may face when driving, such as weather, traffic, and other factors, and to guarantee that AI systems operate consistently under all circumstances.