Overcoming manufacturing data barriers to leverage AI and drive digital transformation

Alan Kavanagh

Senior Digitalisation Programme Manager, Irish Manufacturing Research

Many organisations see AI as a catalyst for transforming their competitiveness. Manufacturers in particular are looking to position themselves to avoid missing out.

Across manufacturing, many companies are in the early stages of their ‘digital transformation’ journey. Instead of being seen as an elixir, efforts in AI should form part of a bespoke, agile digital transformation strategy.

AI investment challenge

Despite a marked increase in AI initiatives, return on investment is mixed. Generative AI, for example, is ranked lowest for deployed solutions and second-last for return on investment.¹ One major factor in the lacklustre performance is the need for most AI solutions to access large quantities of specially prepared data to provide reliable outputs.

Many manufacturers simply do not have enough data ready for AI to use to make meaningful predictions. An astounding 95% of manufacturing organisations still use paper-based processes, with 50% using manual spreadsheets for most of their operations.² More companies lack the digital skills, tools and broader digital infrastructure. The path to realise AI’s potential is therefore coupled with an organisation’s broader digital transformation strategy.

The manufacturing industry is a data powerhouse

The good news is that manufacturing sites are incredibly data-rich places. There is more manufacturing data potentially available to be utilised by AI than any other sector. A typical equipment PLC controller on the factory floor generates a significant 100GB of data each year, while data from machine sensors are the fastest growing subset of all data worldwide. With tens of millions of PLCs integrated into production lines globally, AI has the potential to achieve extraordinary advancements for manufacturing organisations.

The path to realise AI’s potential is coupled with an organisation’s broader digital transformation strategy.

The major challenge is this data can be siloed across a fragmented network of equipment, systems databases, etc. These can be difficult to access, utilise, integrate, scale and secure. These industrial data are often left untouched for many (often legitimate) reasons, such as cybersecurity and IP protection.

The opportunity cost of keeping this conservative approach is quickly rising. Understanding how to safely and efficiently collect and prepare data for AI to tackle manufacturing problems — such as predicting equipment failures, improving output or anticipating supply chain shortages — can help the technology meet expectations.

Industrial data opportunity

Without a digital transformation strategy and infrastructure in place, AI success will be confined to limited use cases. The inflexion point in successful AI implementation will come when more of the businesses’ own data can be readily used for making faster, smarter decisions to solve real problems that deliver real change.