25 September 2023
How can generative AI enable productivity and operational enhancements for the manufacturing industry?
The manufacturing industry has long been at the forefront of large data analytics and machine learning applications. Such tools have been used to gain insights into process operations to optimise across capacity, scheduling, efficiency, maintenance regimes and defect detection. Measured against traditional machine learning, generative AI has the ability to produce new data sets and so produce a fundamentally new level of insights that can unlock a step change in productivity and efficiency enhancements.
We asked our panel of over 400 middle market business leaders about how they felt generative AI would impact their organisations and operations. 77% believe that generative AI could be used to improve their business. Due to mass data generation and potential for smart integrated processes, the manufacturing industry is uniquely positioned to exploit the benefits of generative AI.
Productivity and operational improvements
60% of our panel of middle market business leaders believe that generative AI will lead to increased productivity. This is most likely to come from a revolution to data driven decision making which results from pairing generative AI with process infrastructure, such as real time sensors, process control integrations and digital twinning.
Process optimisation and automation
Systematically integrated processes combined with real time sensor data can utilise generative AI to introduce a new degree of process automation and control. Through the analysis of live data feeds, generative AI can predict bottle necks, optimise material and energy usage, manage production rates accurately and optimise staffing placement. Generative AI can identify new process designs and conditions to enhance each of these areas. When combined with integrated instrumentation and control systems this can lead to optimised production processes, which minimise cost, consumable and utility usage. These improvement will reduce the overall cost of operations and increase efficiency. BMW recently partnered with Zapata Computing, and MIT’s Centre for Quantum Engineering (CQE) to showcase the potential of quantum inspired generative AI techniques to optimise vehicle production schedules across multiple manufacturing plants, this has the potential to substantially increase global capacity and reduce production costs.
To date, predictive maintenance has been dependent on retrospective analysis of equipment performance and manual inspection, combined with real time sensor data feeding into predictive models that are dependent on readily deducible failure modes. Generative AI has the potential to revolutionise predictive maintenance through undertaking advanced behavioural analysis to define and monitor operational windows bespoke to each application more accurately and generate potential new maintenance regimes and failure modes.
Generative AI can analyse data to provide new insights through identifying minute failure modes and make connections between system interactions that would be overlooked by basic linear models and human analysis. This has the potential for issues to be identified before they escalate to a more significant failure, ultimately enabling real time intervention, reducing downtime, optimising maintenance cycles and reducing cost.
By utilising a range of generative AI powered visual systems it is possible to appreciably improve the automation of quality control and defect identification. Generative AI systems can be trained to first pre-process and automatically remove noise and then identify defects, generating new failure modes never previously recognised. This can be done on a macro level and combined with other technologies such as infrared systems to identify defects undetectable to the human eye. Such systems operate on a real time basis and, when integrated with wider process controls, can automate the real time realignment of processes.
Generative AI can analyse a range of live data sensor feeds and identify pathways to safety incidents such as dust environments, leaks and spark initiators. In addition to retrospectively analysing process data to conduct root cause and failure mode analysis. Generative AI can do this in a manner that generates new pathways to failures, which involve the complex interactions of multiple parameters overlooked by a human operative. This enables real time alert of potential incident pathways to enable early intervention and prevention.
To fully harness the power of generative AI it is necessary to fully considered the associated risk of the technology and to build mitigations from an early stage. 63% of our panel of middle market business leaders believe that generative AI will to at least some extent be a potential threat to their business.
There has been significant speculation about the potential for jobs displacement due to generative AI. However, 61% of our panel of middle market business leaders believe that generative AI will lead to some increase in headcount within their organisations.
Generative AI will fundamentally change the workforce of the future for the manufacturing industry, demanding new skill sets and organisational structures. This will impact all corners of the sector from the operational tasks of technicians to disrupting the traditional design and process development approach of engineering technical specialists and impacting the structure of core business functions.
Real time investment
The potential and accuracy of generative AI integration within the manufacturing sector is wholly dependent on the availability and accuracy of data. To unlock the full potential of generative AI integration within the production environment, it will require significant investment and redesign of the factory floor including control and systems integration, real time sensors, digital twinning, and data management and storage.
Transparency and auditing of generative AI systems will be critical to identify bias, inaccuracies, and misleading outputs. This also requires accurate and representative data being made available to generative AI. Consideration will be required to the correct position of generative AI within the decision-making framework to ensure ethical considerations are fully addressed from the offset. Due to the large volume of proprietary data that underpins generative AI systems, appropriate consideration to cyber security issues must also be given.
Generative AI promises the potential of significant innovation, efficiency improvements and cost reductions, however its implementation must be carefully planned with early full consideration of the risks posed.