Exploring the trends, challenges and opportunities of Smart Manufacturing

The Manufacturer’s Henry Anson recently led a virtual roundtable discussion on the trends, challenges and opportunities of Smart Manufacturing, particularly the role of Artificial Intelligence (AI) in industry.

The session — delivered in association with Cognizant and Microsoft — saw a group of senior manufacturing leaders come together to share their invaluable experiences and perspectives on what is possibly the manufacturing sector’s hottest topic right now.

Here are some of the insights from the roundtable, which was held under the Chatham House Rule to encourage openness and the sharing of information by providing anonymity to those involved.

Making more sustainable choices

The conversation was kicked off with a question on how digital technologies can help engineering and manufacturing teams make more sustainable choices, potentially through taking advantage of supply chain data to boost efficiency.

The discussion focused on the implementation and integration of digital technologies and data management in industry, particularly through the lens of digital thread and digital twin concepts. Participants highlighted the crucial role of robust data engineering in integrating disparate data sources, such as those from supply chains, manufacturing processes and design departments. There was an emphasis on creating data models that not only support complex analyses for data scientists but also offer simpler, accessible views for business users to facilitate quick decision-making.

Potential applications of AI and machine learning were explored, especially for alerting users to significant data trends like process deviations or anomalies. Although there is interest in taking advantage of these advanced technologies, current efforts are still in the early stages, with a primary focus on consolidating datasets and tools to enhance overall accessibility.

The group also delved into technology infrastructure, comparing on-premises systems with cloud-based solutions. There was a specific mention of transitioning to Microsoft systems, highlighting the necessity for a solid backbone for data integration. This backbone would ideally consist of ERP, MES and design information systems like Teamcenter. However, despite these advancements, challenges persist in data sharing with customers and suppliers, often due to security concerns and a continued reliance on traditional paper or file-based systems.

Advice on piloting and implementation strategies was shared, emphasising the importance of addressing specific business problems rather than adopting technology for its own sake. Participants warned against conducting pilots only in ideal conditions, stressing the need to test in varied environments to ensure broader applicability and success.

Cost considerations were a significant point of discussion, particularly the high costs associated with indiscriminate data collection and storage. The consensus was that specific business needs should drive data collection efforts to avoid unnecessary expenses. Additionally, the importance of incorporating technical risk assessments early in the design and manufacturing process was underscored, as this can streamline concept selection and improve overall agility and effectiveness.

The opportunities presented by AI in industry

The discussion then moved on to the significant advantages of AI in various industries, noting that almost every business, regardless of industry vertical, is beginning to see tangible benefits. The conversation referenced an industry survey, which revealed that cloud and AI are the top technologies driving smart manufacturing. Participants identified three levels of AI benefits: individual convenience, enterprise efficiency and societal impact. Examples included AI-driven shopping recommendations and streaming services for individuals, while enterprises utilised AI for operational efficiencies, such as vision AI for quality control in automotive manufacturing.

The societal benefits were exemplified by the NHS’s use of AI for disease detection and Network Rail’s predictive maintenance. Network Rail employs high-speed cameras on locomotives to detect maintenance needs and enhance safety by identifying track obstructions. These examples underscore AI’s broad applicability and significant impact on various operational aspects.

Moreover, the UK is recognised as a favourable environment for AI development. Notable AI companies like OpenAI, Microsoft and DeepMind have established hubs in the UK, driven by the country’s strong academic support and government investments, such as the recent £7.4m funding for AI training and a partnership with the US on AI safety. The UK’s lead in AI company setups within Europe was also noted, indicating a robust AI ecosystem.

However, challenges persist, particularly in competing globally. One participant highlighted Europe’s higher energy costs compared to China and the potential of AI to improve energy efficiency and sustainability. AI could help reduce energy costs by dynamically responding to electricity prices and optimising renewable energy use. Additionally, AI’s role in ensuring traceability of raw materials and sustainability in production could offer competitive advantages.

In the space industry, another discussion participant noted the challenge of small datasets and the early stage of AI adoption. Despite this, AI is used to scan solar arrays for faults and support innovative applications using satellite data. The potential lies in merging satellite data with other datasets to uncover valuable insights.

Lastly, someone pointed out that AI in the Maintenance, Repair, and Overhaul (MRO) sector focuses on improving local competitiveness rather than competing internationally. AI applications in MRO include data extraction from emails and PDFs to accelerate response times to customers. Overall, the discussion emphasised AI’s transformative potential across different industries and the UK’s strategic position in AI innovation.

Predicting what the road to Net Zero will look like

In the volatile automotive market, predicting the future of vehicle technology poses a significant challenge. An individual from a British multinational chemicals company, with over 200 years of experience in internal combustion and other technologies, highlighted the uncertainty associated with achieving Net Zero. “Batteries alone aren’t a complete solution,” they said and while hydrogen is promising, it’s costly. Amidst the political lobbying and critical debates, AI could offer a pathway to better predictions, thereby aiding companies and their customers and suppliers alike in making more efficient decisions.

Responding to this point, another participant noted the debate on battery production, mentioning that EVs require six times more minerals than traditional vehicles, merely shifting pollution across the value chain. Despite the political debates, AI’s potential lies in enhancing R&D and engineering. AI has already accelerated advancements in areas like new molecule discovery, notably illustrated by breakthroughs in protein folding for drug discovery. One participant outlined how their business collaborates with AI firms to develop better catalysts, showcasing the transformative impact of AI in targeted research.

The conversation then turned to broader AI applications in the automotive industry, suggesting that AI can optimise the design and efficiency of competing technologies, such as hydrogen fuel cells. For instance, AI can handle vast permutations and combinations to arrive at optimal designs quickly, significantly cutting down research time. An example was provided where AI reduced 38 million material combinations to 125 viable options using quantum computing, underscoring AI’s power in expediting R&D processes.

It was pointed out that while AI excels in isolated R&D tasks, it should also address larger market predictions. Currently, market forecasts, like the 45% penetration of battery electric vehicles, are based on assumptions. Given the fluctuating economics of lithium and potential environmental impacts, these predictions are uncertain. AI could help refine these forecasts by integrating various dynamic factors, thereby providing a more accurate market outlook and aiding in strategic decision-making.

However, another participant cautioned that while AI can help identify new materials and technologies, its predictions can be speculative. Historical examples, such as the evolution of nuclear materials from plutonium to thorium, illustrate how AI-driven research can pivot market trajectories. Thus, AI’s primary value lies in facilitating R&D and engineering, potentially reshaping technology landscapes through innovative discoveries.

As the discussion concluded, it was emphasised that AI’s utility depends on clear objectives and careful design of algorithms. AI’s responses are only as good as the prompts it receives. Properly scoped and directed, AI can combine data from myriad sources to inform decision-making, reducing the burden on human computation. However, the potential for bias — both from the algorithm and from human prompts — necessitates a collaborative approach to AI-guided insights, ensuring balanced and accurate outcomes.

The potential role of AI for H&S

Participants then explored the potential and ethical implications of using AI for workplace safety. One individual emphasised that AI should serve as a guidance tool rather than an autonomous decision-maker. A key suggestion was leveraging AI to analyse existing CCTV footage to ensure workers are wearing appropriate Personal Protective Equipment (PPE). This could enhance safety by triggering alarms if PPE compliance is not met, particularly beneficial during high-risk activities like maintenance or chemical handling.

This led to an important ethical consideration being raised about using AI for monitoring in workplaces, questioning how unions might react to such surveillance. However, it was clarified that the AI application discussed involves analysing general surveillance footage rather than tracking individuals. The focus is on identifying non-compliance with safety standards without infringing on personal privacy, thus addressing some ethical concerns.

Another participant noted the potential union challenges in implementing AI surveillance, citing their experience at a British aerospace and defence company. Despite recognising the value of AI in improving safety, navigating union agreements and worker consent would be complex. This highlights the need for careful consideration and negotiation when introducing AI monitoring in unionised environments.

Towards the end, an additional suggestion was made about using wearables as an alternative to cameras for monitoring exposure to hazardous materials. This could complement vision AI and provide more direct and personal safety monitoring. The discussion underscored the balance needed between enhancing safety through technology and addressing ethical, privacy and union-related challenges.

Finding the right skills

The roundtable was concluded with a discussion on the challenges of recruiting and training individuals with the right combination of skills for modern manufacturing roles, which increasingly require expertise in both traditional engineering disciplines and digital technologies. One participant highlighted the difficulty of finding individuals who possess skills in mechanical design, software coding and systems understanding. The consensus was that this problem is widespread and affects many in the industry.

Another speaker shared that they have expanded their engineering team significantly but found it challenging to find so-called “unicorns” with the perfect skill set. Instead, they fostered a collaborative team environment where individuals could learn from one another and develop multiple skills. This approach helps bridge the gaps in specific skill sets through teamwork and continuous learning.

In Germany, the situation is similar, despite a higher regard for manufacturing, according to one participant. The German manufacturing sector, like its counterparts in the UK, they said, faces similar challenges in integrating new technologies and finding multi-skilled workers. The discussion touched on the competitive threat posed by countries like China, which has a vast workforce of engineers dedicated to advancing these technologies.

One solution proposed was to focus on hiring individuals with a willingness to learn and grow within the company. Graduate programmes and robust training models can help new hires develop the necessary skills. Emphasising soft skills, such as adaptability and motivation, is crucial given the rapid pace of technological change in manufacturing.

Several participants mentioned the importance of internal training programmes. It was highlighted how some companies have learning academies to prepare for future skill needs, while others implement apprenticeship programmes that combine practical experience with digital skills training. Additionally, other companies are leveraging partnerships with educational institutions and service providers to access training materials and expertise.

The discussion underscored the importance of creating a culture of continuous learning and adaptability within organisations. By investing in employee development and fostering an environment where employees can learn new skills, companies can better navigate the evolving landscape of manufacturing and maintain a competitive edge.

Final thoughts

As we’ve heard, Smart Manufacturing powered by AI presents a transformative opportunity for businesses to achieve substantial operational enhancements. From predictive maintenance reducing downtime to AI-driven analytics optimising production processes, the potential for real business improvements is profound.

Embracing these technologies not only enhances efficiency and quality but also enables agile responses to market demands. However, effective implementation requires robust data privacy measures and strategic alignment with business goals. By harnessing AI’s capabilities, businesses can not only streamline operations but also foster innovation, ensuring they remain competitive in an increasingly digital landscape. Smart Manufacturing stands poised as a cornerstone of future industrial success.

Thank you to Cognizant and Microsoft for enabling the discussion, and to the senior manufacturing leaders who gave up their time to participate and share their insights; these extremely valuable roundtables are not possible without you.