Investment in artificial intelligence (AI) for manufacturing is expected to grow by 57 per cent by 2026, from US$1.1 billion in 2020 to US$16.7 billion by 2026. This is a remarkable leap forward that will see machines mimic aspects of human intelligence and automate complex tasks.
This could not only boost revenue and lower costs, it could also reduce risk and transform manufacturing.
Embracing human-machine collaboration now will allow industries to augment human capabilities and introduce great new possibilities for using AI to transform manufacturing.
But most manufacturers – 56 per cent – are still only using AI in small-scale pilot projects, according to the Manufacturing Leadership Council 2023 survey, while only 28 per cent have passed the pilot stage.
Advanced manufacturers such as producers of industrial machinery or aircraft components are already adopting AI to deliver more intelligent manufacturing. This is not just about using new technology, it also involves company culture, empowered people and net zero commitments.
Other manufacturers could look to the experiences of these first movers to see how they’ve successfully tackled the challenges of using AI to transform manufacturing. These four examples show how scientific instrument manufacturer Agilent Technologies is using AI to transform manufacturing:
To help bring science to life, Agilent manufactures vital scientific and research instruments. We rely heavily on manual work to test our products. When inefficiencies in the production process caused us costly delays, we built a dedicated digital solutions team to design AI-based predictive and data-driven testing.
We deployed 250 industrial Internet of things (IIoT) stations that use AI algorithms to learn from previous test results, identify patterns and maintain automated tests. This helped us to shorten product testing time, improving the length of our work cycles by 23 per cent.
Customer expectations rise as technology advances and the assembly work of scientific instruments becomes much more complex. Normally, this increases labour costs and lowers efficiency, but using AI can increase flexibility when responding to changing customer demand. AI can support root cause analysis, natural language processing and time series modeling to identify the underlying causes of a quality shortfall faster and more accurately. At Agilent, we’ve used these AI-powered tools to reduce production downtime by 51 per cent.
In response to a 2050 net zero emissions goal, and to ensure the workforce is Fourth Industrial Revolution (4IR)-ready, Agilent has a dedicated team that develops big data architecture and analytics. This multi-disciplinary squad uses AI to simplify test processes without compromising product quality.
Advanced machine learning algorithms help to rapidly analyse and identify ways to improve testing processes. As a result, Agilent has reduced recycled waste by 53 per cent and improved productivity by 31 per cent.
New market dynamics and technology trends are driving the need to future-proof highly complex manufacturing operations. To do this, Agilent has implemented the "lights-out factory" – a fully automated technology that runs product testing “in the dark”, that is with no human intervention.
This is enabled by a company culture focused on trust, respect and innovation, alongside training and development to help reshape human-machine collaboration.
A lights-out factory is not easy to implement in complex manufacturing. Agilent used robotics process automation and AI technologies to achieve sustainable production and overcome the bottlenecks that can arise with such a set-up.
Central to this is a comprehensive, harmonised IIoT with real-time instrument sensing for indoor positioning. Overall productivity has increased by 33 per cent as a result.
There are three initial steps that all complex manufacturing companies should consider when preparing for an AI-powered future:
Leadership commitment and investment is crucial for companies using AI to transform manufacturing.
Leaders must understand how to scale and be realistic about the timeline from concept to execution. This begins with a people-first mindset, building workforce capabilities and starting with small-scale pilot projects that deliver tangible outcomes.
Selecting the right process to change lays the foundation for successful AI integration and operational improvements in manufacturing operations. Consider processes that require high-level attention, such as those that depend on manual work for decision-making.
Any operational performance that has a high impact on your organisation, or manual methods that cannot easily detect anomalies and deviations, are also good places to start when thinking about how to integrate AI.
Careful consideration of technology platforms is pivotal. Start with a technology platform based on the existing strengths of your organisation. Build on your team’s AI capabilities before deciding on the right tools.
This article was first published by the World Economic Forum.
