AI propels smart manufacturing towards a sustainable future

This article was originally published in Dutch by ITExecutive. It has been translated into English below.

The rise of the intelligent factory: Efficiency and sustainability converge

Manufacturing is undergoing a metamorphosis- mere automation is no longer enough. Manufacturers are struggling under relentless tensions: stringent regulations, shorter product lifecycles with consumer demands for sustainability, supply chain disruptions, and ever-present cost pressures. Traditional manufacturing is struggling to keep pace and even "smart factories" are becoming table stakes – a steppingstone in the face of this disruption. A manufacturing renaissance is needed, and artificial intelligence (AI), specifically generative AI (GenAI), can be the accelerant propelling smart factories become more adaptive towards an intelligent and sustainable future.

We sat down with Hemant Singhal and Suman Kumar Sengupta from Cognizant, both specialists in the intersection of operational tech and smart solutions, to find out how AI can help the manufacturing industry.
 

Beyond automation and efficiency: AI ushers in a new era

While the core principles of manufacturing remain constant, "smart manufacturing" represents a paradigm shift due to powerful convergence of technologies. Real/near-real-time data streams from Internet of Things (IoT) sensors and devices on machines, factories and products, enterprise data (such as inventory), external data (such as weather or market) on zero/low latency networks form the backbone of this intelligent ecosystem. AI, machine learning (ML), and big data analytics use this data backbone to intelligently optimize production, prioritize safety and sustainability, and ultimately, boost revenue growth.
 

Generative AI: The disruptive catalyst for sustainable innovation

Traditional AI is good, but GenAI is game-changing. GenAI doesn't just analyze data; it invents entirely novel designs and solutions. Imagine sustainable product designs not even conceived yet, adaptive inspection criteria (for maintenance schedules) due to ever-changing environments for proactive maintenance, minimized downtime and optimized resource allocation.
 

Current state of adoption

Manufacturing companies are already experimenting with GenAI, setting up dedicated teams, executing tangible use cases, and building proofs of concept (PoCs). “It won’t be all that long before GenAI leaves this formative stage and takes off completely”, is the expectation of Singhal and Sengupta. At the same time, traditional AI can’t be ignored and GenAI must complement it, so they synergistically drive intelligent and sustainable manufacturing.
 

The future of work: Humans and AI – Partners, not rivals

This whitepaper, authored by Cognizant, examines how AI and GenAI can demonstrably enhance manufacturing processes and sustainability initiatives. By embracing AI as a partner, not a rival, investing in skills development, fostering collaboration, and prioritizing ethical development, we can create a future where AI complements and augments human capabilities, unlocking a new era of productivity, creativity, and shared prosperity.

An AI-powered assistant for technicians is one example of a tangible use case. "A technician can ask questions and based on available documentation and knowledge base based on past incidents, the assistant can give possible solutions for a given problem through an augmented reality interface. Imagine a technician receiving real-time solutions overlaid directly onto their field of vision through an augmented reality interface”, says Singhal. "Technicians won’t have to work their way through a 1,000-page manual at 2:00 AM anymore."

"Another important aspect here is that an innovation like an AI assistant helps make the conversation on AI more human," adds Sengupta. "These kinds of features are helping the blue-collar community to adopt an otherwise very complicated AI environment."

A dissimilar, but equally important, use case for GenAI in manufacturing is personalized training recommendations for workers. An HR manager can have an AI agent generate a training plan tailored to any person’s role, needs, or knowledge.

In the future, AI agents, humans and robots will team together for a safe, high-performance, and rewarding work environment.
 

Digital twins and GenAI

Manufacturing companies are facing immense pressure to develop superior products at a faster rate, while also adhering to sustainability measures and reducing costs; both the regulatory environment and consumers demand responsible sourcing and production. "Digital twins are part of the answer because instead of creating something that fails, which takes a lot of time, you first create a digital model of whatever is going to happen on the shop floor," says Singhal.

Digital twins – virtual representations of physical objects – offer a powerful solution. However, traditional digital twins rely on historical data, limiting their design potential.

GenAI shatters these limitations. By possessing the ability to "think outside the box", GenAI can generate entirely new designs optimized for sustainability. Imagine creating products with minimal environmental impact—surpassing even human imagination!

"With the creation of a digital twin, you can also give constraints to the model," says Singhal. "You can say: 'Generate this product with the minimum amount of energy, water or input material', so you create more sustainable & efficient products."
 

ESG reporting and responsible production: Transparency redefined

"The case for generative AI enabling smart manufacturing is very clear," Singhal says. It can assist in responsible sourcing and supply chain transparency: it can generate reports on how you source raw materials, and how you transport and use them. It can also create reports from a risk-management perspective. All of this is very helpful for reporting and governance requirements.”

You could refer to this kind of solution as an AI agent. "Almost like a human being, you give the agent an objective, such as 'meeting ESG goals’, and the AI will orchestrate the entire supply chain and production ecosystem to meet that objective. It will monitor whatever is needed and take action while keeping humans in the loop."
 

Limitations

GenAI has a lot of potential for companies in manufacturing, but there are valid concerns, according to Sengupta. "We see a lot of hesitation among customers. Nothing is in production yet, and everything they do is on a small scale. There’s a catch-22 situation: if experiments are not at a certain scale, you don’t see the value yet."

There is also another concern, he tells us. GenAI is still in its infancy, and businesses don’t have architectural standards in place yet. "What should a CIO use, an open-source model or a proprietary model? Is it production-ready? Is it at a scale they want to use, and is it secure enough?" This is the reason that adoption is delayed, according to Sengupta.

Then there’s the people aspect. "At the corporate level, they are convinced that GenAI will help people and augment their safety. But they need to manage the expectations of the people on the shop floor: plant managers, engineers, office workers, production leaders, etc. Accelerated deployment can run into resistance from these personas- is AI really going to help them? Or is it going to replace them?” Addressing such concerns are crucial.
 

From innovation to implementation: The generative AI journey

So how do you make GenAI in manufacturing a reality? How do you industrialize AI projects? "What we want to do, is use the experience we have with traditional AI and apply some of the lessons learned to GenAI as well," Singhal says.

According to him, and contrary to popular perception, many AI/ML solutions do not enter production not due to lack of data scientists or technology, but because of data issues. "If you build a POC with a skewed or unrepresentative sample, or bad data quality, you get results that are too-good-to-be-true, trivial or simply unreliable." This has led to boards unwilling to invest further in new projects.
 

So how do you get started?

• Ditch the pilot paralysis
  Don't get bogged down in endless proofs of concept. Show value by executing high-impact business use cases such as sustainable product design or predictive maintenance, and then aggressively scale.
  
  
• Embrace the data imperative
  GenAI is data-thirsty. Develop a robust AI strategy underpinned by robust, scalable data strategy. Build a data literate and data-driven culture across your organization.
  
  
• Augment and reskill, don't replace
  The human element remains vital. Upskill your workforce to leverage AI assistants, collaborative robots (cobots) and make them active partners in the AI revolution.
  
  
• AI center of excellence (CoE)
  AI success needs a multi-pronged attack. A strong governance body, like an AI CoE, is your secret weapon to drive a multi-disciplinary approach combining data, AI, software, and platforms while aligning business vision with AI strategy, fostering a data-driven culture and breaking down silos.

The time for incremental change is over. GenAI presents a transformative opportunity to unlock a future where manufacturing thrives alongside a healthy planet. Are you a leader, or a follower?