Generative AI in Manufacturing : A transformational alliance

Artificial intelligence (AI), in particular, Generative AI is emerging in various applications in retail and customer-facing businesses. Wherever extensive data is already available in the form of text, speech or images, Large Language Model (LLM) open-source software resources are propelling newer applications on a daily basis to enhance customer experience, thereby providing a competitive edge to businesses. We see more Generative AI applications - using ChatGPT/BARD - in content writing, contextual translation, code refactoring, and customer service applications in retail, e-commerce, IT/ITES service, edtech and fintech industries.

The manufacturing industry in Bharat can attempt to deploy basic AI / ML techniques far more creatively than it does today. AI / ML is generally considered as a 'black box' for two specific reasons:

1. Tacit data dominates, i.e. the people on the shop floor somehow still fear data transparency and automation will lead to job losses

2. Data collection and integration is complex owing to the existence of a multi-product, process, and equipment architecture with multilevel silos of 'protected architecture' of several 'proprietary brands'.

However, Gen AI can be a transformational partner for the manufacturing industry. And give it the cutting-edge it needs.

Also Read: Building AGI stack to shield against potential frauds, says Paytm CEO

As an example, in an electronic manufacturing company, in order to debug a malfunctioning Programmable Logic Control (PLC), in a Made in China equipment, the shop floor maintenance engineers needed the equipment manual to be translated to English as it was written in Mandarin. In addition, they were also required to take the assistance of a Chinese engineer through a video call. It was a tedious process, and valuable time was lost. The issue could have been addressed instantaneously with the deployment of Generative AI, at least for all critical machines in the company, along with augmented reality & virtual reality (AR / VR) technology. 'Experts' for 'black boxes', currently needed for firefighting, can be far more effectively utilised for preventing fires and many other predictive and process optimisation opportunities.

Potential application areas of Generative AI in the manufacturing industry could be:

1.    Product design and development

2.    Predictive maintenance

3.    Quality control and assurance

4.    Production planning and inventory management

5.    Supply chain management

These applications help not only to bring innovative products based on cutting-edge technology to the market but also provide significant opportunities in improving cost, quality, efficiency and overall competitiveness.

As an example, 'Art to Part' in product design and 'Part to Art' when reverse engineered. In both scenarios, the manufacturing industry can use Generative AI. A couple of examples of product design with Generative AI applications are : (i) the Bionic Eye i.e. assisting the visually challenged to 'see'; and  (ii) new generation compact PCB architecture with flawless embedded design and software for Electric Vehicle battery optimisation.

Another interesting example is using CCTV video analytics in real-time for safety, health & ergonomics on the shop floor. Further, machine vision and camera inspection techniques remove subjective inferences and conclusions from quality control inspections on the shop floor, thereby providing greater quality assurance. Also, real-time algorithms from retail success stories on demand pattern analysis can help in production planning and inventory management methodologies. The same is the case with opportunities for Generative AI applications in supply chain management, the smooth functioning of which is more critical today than ever before for the manufacturing industry.

Apart from the shop floor, all the other functions like sales & and marketing - new proposal submission for complex customer deals; finance - multiple fintech case studies can be replicated; HR - recruitment and talent management, especially the induction training for new entrants,   can effectively apply Generative AI. In addition, B2B clusters in the MSME sector, such as foundry or, machining or injection moulding, can leverage many improvement opportunities together. Examples like milk run transportation and logistics, non-moving inventory liquidation, collective purchase of raw material, yield improvement, working capital financing, etc come to mind immediately.

Also Read: AI detects gallbladder cancer as accurately as radiologists in India: Study

Fortunately, almost all the tech giants - like Microsoft, IBM, SAP, AWS and Google - whose software products are already in some way or another in use in the manufacturing industry, have made massive investments in Generative AI. What the industry needs to do is to train and re-skill its employees in AI / ML concepts and applications. In fact, most software products are now in 'No Code' mode and bring the necessary 'glamour quotient' for attracting graduate engineers. The traditional statistical experts in TQM, TPM, TCM, Lean, Six Sigma should now move rapidly to embrace emerging technologies to accelerate the growth and profitability of the manufacturing industry.

The existing data on the manufacturing shop floor, in whichever form it is - text, image or speech or from whatever source - ERP, MES, HRMS, CCTV footage, Quality Management, Occupational Health & Safety systems etc, can be resourcefully leveraged to derive patterns, create new set of futuristic information and improvements through Generative AI. This will change the perspective we mentioned in this article's opening statement about the 'black box' nature of data transparency and interoperability of rigid architecture in the manufacturing industry. The latest Industrial IoT hardware and software architecture is

predominantly 'brand' agnostic and can, therefore, greatly assist in breaking down the silos created by proprietary products in data collection.

Further,  we quote from Prof C K Prahlad's book, "The New Age of Innovation," where he proposes an empirical formula, 'N=1' and 'R=G' emphasising that there are new managerial demands in business, requiring new sources of value creation. He argues that these demands have created an N=1 and R=G environment, where companies need to customise their product for each unique customer by gaining access to a new array of global resources. Generative AI brings ample opportunities of open resources for deploying hyper-personalised concepts for new technology products and processes in the manufacturing industry.

There is a great opportunity for manufacturing companies individually and the manufacturing industry as a whole in Bharat in its growth aspirations to reach a global scale, taken together with Government support through the PLI schemes and the geo-political situation leading to a China+1 strategy of global corporations. In fact, AI in general and  Generative AI in particular, have the capability to form a transformative alliance with the manufacturing industry to provide it with the cutting-edge it needs. This adoption of state-of-the-art technology will also attract new-age talent from schools, colleges and universities towards the manufacturing industry, which has been under the shadow of the service industry for the last few decades. Our manufacturing industry would do well to consider exploring the opportunity that AI and Generative AI provide.

Vipin Sondhi, Former MD & CEO of Ashok Leyland and JCB India and G. Sundararaman, Co-CEO Wipro Pari