Professor Denis Dowling
Director, I-Form
Manufacturing companies can leverage AI to gain a competitive advantage. However, they also face various AI adoption challenges that must be overcome with careful planning.
Professor Denis Dowling, Director of I-Form, the Research Ireland Centre for Advanced Manufacturing, headquartered at University College Dublin, works with industry partners on bespoke solutions that make their processes more efficient, sustainable and profitable.
Crucial advantages for manufacturers adopting AI
He outlined that AI presents manufacturers with numerous important advantages. “Take predictive maintenance,” he says. “Sensors and data analysis can help identify problems in the manufacturing process at an early stage. Manufacturers can then modify the process or stop the equipment, but either way, they will have enhanced processing efficiency, because they won’t be wasting their materials and energy.”
For example, Prof Dowling recalls how I-Form developed a bespoke machine learning-based Recommender System for one SME partner. This analyses sensor data in real-time to detect manufacturing defects and then sends recommendations to an operator’s mobile in under nine seconds. The SME was able to enhance the yield from its manufacturing process, while reducing both waste and development costs.
AI also offers manufacturers valuable predictive modelling capabilities. For instance, when I-Form developed a set of bespoke AI-based simulation tools for the design and manufacture of engineering components, the company that trialled the tools reported a 90% reduction in the number of physical experiments needed to develop its new product line.
If poor information goes in, then poor information comes out
A careful and strategic approach for successful AI adoption
Nevertheless, manufacturers should be aware that successful AI adoption must be carried out carefully and strategically to overcome the various challenges they will inevitably confront. These include high upfront investment costs, integration difficulties with legacy systems and a lack of skilled talent. Access to good-quality data is also key. “If poor information goes in, then poor information comes out,” says Prof Dowling.
Cultural resistance from employees is another issue. “There’s a human-centred side to AI adoption — including the upskilling of staff and job changes — which is important to consider,” says Prof Dowling. One way to make AI adoption easier is for manufacturers to engage in collaborative research with academia. They can then access cutting-edge knowledge, specialised expertise and advanced facilities without heavy upfront costs.
Partnerships also support long-term skills development, attract funding opportunities and ensure alignment with ethical, regulatory and sustainability standards, making AI adoption in plants more practical and scalable.
Those that successfully adopt AI have the potential to use it to enhance efficiencies,” says Prof Dowling. “Those that don’t do so are likely to be at a disadvantage.”
