Manufacturers are finally escaping AI ‘pilot purgatory', as they turn to disruptive technologies in order to take on the challenges presented by COVID-19. According to new research by Google Cloud - which polled more than 1,000 senior manufacturing executives across seven countries - 76% of organizations in the sector have adopted data analytics, cloud and AI during the pandemic.
This is in stark contrast to 2020 findings from Gartner, which outlined that only 21% of companies in the industry had active AI initiatives in production. Google Cloud has found that almost two thirds of manufacturers (64%) now rely on AI to assist in day-to-day operations, with a quarter already allocating half or more of their overall IT spend towards AI.
Google Cloud's upcoming Digital Manufacturer Summit will see Google executives and customers discussing how organizations in the industry are embracing cloud and data tools to boost productivity, reduce costs and improve quality. You can register for the event here and access all of diginomica's coverage in our dedicated events hub here.
Dominik Wee, Google Cloud's Managing Director of Global Manufacturing, Industrial and Transportation, told diginomica that the technology capabilities that are now available to manufacturers mean that they are now able to scale out their AI pilots, making use cases repeatable. He said:
I think manufacturing is traditionally a very conservative industry. I think most manufacturers or plant floor managers will always trade off continuing the factory running, over doing innovation. Every change is a risk. The more disruptive these technologies are, the more hesitation there is - and for very good reasons.
But I think what's happened through COVID-19, in many ways, is that it's taken some of this hesitation away and has also created a necessity to adopt. I think the pandemic has opened minds and I feel like there's a lot more openness to our technology than previously.
This openness has largely been driven by manufacturers becoming painfully aware of their supply chain risk during the pandemic, with visibility into how products are sourced, and where they are at any point in time, being critical to supporting operations.
Equally, in order to gain more control, manufacturers are increasingly turning to insourcing and reshoring operations. This has meant a renewed focus on automation, as there is an entirely different labor factor cost.
Diving into the data
According to the Google Cloud research, manufacturers are turning to the use of AI in day-to-day operations for a number of specific reasons. These include: assistance with business continuity (38%); helping make employees more efficient (38%); and to be helpful to employees overall (34%).
In terms of the use cases that Google Cloud is seeing in the industry, manufacturers are seeking support with quality inspection (39%), supply chain management (36%), risk management (36%), product and/or production line quality checks (35%), as well as inventory management (34%).
However, Google Cloud is also seeing use cases in areas that include powering connected factories, to assisting with predictive maintenance.
Despite the increase in adoption of AI, data tools and cloud, Wee does note that when manufacturers struggle to move beyond pilot stage, they face three key barriers. Buyers should understand these and formulate plans to overcome them in order to see the best results. Wee said:
The barriers to adoption are threefold. One is around the availability of data, particularly when you look at the manufacturing shop floor, data is often very siloed. Even within one factory it might be siloed, but then you may have 100 factories scattered around the world.
The second one is legacy - not just from a data point of view, but understanding the systems and knowing what's where and how to operate stuff.
The third one is talent. I think AI technology in the past has been very hard to use and you needed highly specialized data scientists to run it. That's a gap we can close by making the technology easier to use.
And it's on this last point that Wee and Google Cloud believe that significant progress has been made, where cloud and AI tools have become more accessible and scalable for manufacturers. The Google Cloud research found that having the talent to properly leverage AI was the most cited barrier (25%) by organizations in the industry. Wee explained:
I think what's going to happen is that we are going to see the proliferation of AI use cases. In the past you needed a highly specialized team to solve one problem. If we make it accessible to everybody, and do what Google describes as democratizing AI, you put into the hands of a lot more people.
You start to solve specialized problems that in the past you wouldn't have touched, because the overhead of solving the problem would be too high.
AI use by geography
Looking at the Google Cloud research, it's also interesting to note the variation in use of AI in manufacturing across different countries. For example, while 80% and 79% of manufacturers in Italy and Germany respectively report using AI in day-to-day operations, that percentage plummets in the United States (64%), Japan (50%) and Korea (39%).
The below image illustrates this divergence clearly:
Wee concludes that manufacturers are heading into a ‘golden age of AI', but was keen to highlight two key points. Firstly, whilst manufacturers have been stuck in ‘pilot purgatory', this does not mean that pilots are a bad place to start. He explained:
Some manufacturers may be stuck on pilots, but that doesn't reduce the merit of a pilot in itself - start something meaningful, somewhere meaningful. I think the other most important thing is to start from a business perspective, and not to do it for the sake of technology. You can invest a lot of time and resources into a technical solution that works perfectly in the end, but the financial return is limited. Look at where the cost is and try to focus on that.
Finally, Wee highlighted the important point that whilst AI can reduce costs, improve productivity and boost quality for manufacturers - it isn't being deployed in order to replace technical workers in these organizations. It's being deployed to aid their work. He added:
I think what's important to say that none of these technologies are here to replace the worker. It's not about replacing the work, it's about augmenting what the worker does. What you need to do is take these technically skilled people that you have on the shop floor today and enable them to use the technology.