Artificial intelligence in business, and especially in the manufacturing industry is predicted to hit explosive growth in the next five years - bringing widespread automation, transforming supply chains and building an industrial Internet of Things (IoT). Markets and Markets predicts that AI spend in the global manufacturing market will grow from $1 billion in 2018 to $17.2 billion by 2025.
You might be expecting this article to tell you what to consider before investing in AI. But the numbers suggest that there’s nothing to think about. AI is the next enabling technology for the manufacturing industry, it’s cheaper than you might think and it pays for itself in increased output and better managed supply chains. Investing in the right AI technologies to support your business now will prepare you for the Industry 4.0 revolution.
When you think about AI in manufacturing, the first image that springs to mind is legions of robots in factories performing incredible feats of precision engineering beyond the ability of human beings. But the ‘intelligence’ of those robots is actually pretty narrow.
While they may move better than people, there’s no real thought behind their actions and they often have to operate away from everyone else so they don’t accidentally crush someone. What they’re doing is better classed as automation – basically, the ability to perform repetitive, often intricate, manual tasks (very well). This is a part of AI, but it’s not the whole story.
“There's a lot of confusion around what AI really is,” says Dr Cristian Gherhes, Research Associate at Sheffield University Management School.
“If you call automation AI, then probably more than 90% of what's out there can be labelled as AI. But if it has to improve decision-making and reflect and resemble human intelligence, then it’s probably less than 5% of what’s out there.”
So automation is one important part of AI – a second key element is machine learning algorithms. Machine learning is exactly what it sounds like, the ability to learn from experience (as people do) and improve the way the program carries out tasks.
As machine learning algorithms evolve, the abilities of robots improve too. Robots powered by machine learning can think, to a certain extent, and make their own decisions based on real-time information. These “collaborative and context-aware” robots can use computer vision and AI to work alongside human beings. Instead of being restricted to a single repeating task in one area, for example, they can move around the factory floor and sense and react to people and obstacles in their path.
Intelligent robots are already at work in the FANUC plant outside of Tokyo, where the machines build, test and inspect themselves 24 hours a day. These intelligent robots are building more smart robots just like them, capable of machine learning and computer vision.
Collaborative and context-aware robots will greatly improve production, according to a McKinsey report. Businesses using artificial intelligence can sometimes see productivity increases of up to 20% for certain tasks – even when tasks are not fully automatable.
So shouldn’t you wait for robots to get smarter before you invest? Actually, investing in automation on the factory floor now will help prepare your business for the future. Every step on the technology ladder depends on you having existing technology in place, so factories that are already automated will be able to quickly adapt to smart robots.
But the factory isn’t the only place where automation can boost productivity for manufacturing businesses. Across all industries, AI can automate the most boring and time-consuming business processes using software robots that are frequently offered as part of software-as-a-service (SaaS) packages. Robotic process automation (RPA), as it’s known, automates menial and repetitive data entry and analysis tasks.
Mercedes-Benz uses AI to improve sales processes for its truck division in Brazil, for example. The Microsoft Azure Machine Learning tool uses information, from registration numbers and local legislation to sales information and statistics, to help salespeople chase leads.
Delivering customer experience reports, collecting and processing sales and marketing information, customer-facing chatbots to answer phones – these are all examples of RPA at work. Because these can be featured in vendors’ SaaS offerings, they are also often a very cost-effective first step into adopting AI.
As though robots on the factory floor and in your software wasn’t enough, you also need to put an internet-enabled sensor on pretty much everything. This is the industrial Internet of Things (IoT), which will use connected sensors to gather huge reams of data and process it for new business insights. In manufacturing, we can already see this happening in supply chain management.
According to a McKinsey report on “Supply Chain 4.0”, the potential impact in the next two to three years is huge - up to 30% lower operational costs and a reduction of 75% in lost sales. Sensors on everything from trucks to individual crates are also expected to increase agility in the supply chain while decreasing inventories by up to 75%.
Today, AI is already having a big impact on the supply chain. IBM, for example, says it has shortened managing any disruption down to just hours from 18 to 21 days and halved its expedite costs.
AI can process massive amounts of information and that allows it to see trends and provide solutions faster and more intelligently than traditional software. With the supply chain, that means adding factors you might not usually consider when planning for supplies, like consumer spending habits or the weather, to make highly accurate forecasts about exactly what you’ll need. Armed with that kind of information, manufacturers can improve inventory control, staffing, energy consumption and production and reduce errors that would lead to wasted stock or worker-hours.
There’s another reason that AI has taken off so well in this role. SMEs in the supply chain have a lot of motivation to adopt the latest technologies, because original equipment manufacturers (OEMs) tend to only work with smaller businesses that have the highest quality standards, according to Gherhes.
“SMEs often work together with OEMs in developing new or more efficient processes by employing such technologies,” he says.
“Equally, where an SME has the ambition to work with an OEM, it normally takes steps to enhance the quality of its offering and the efficiency of its operations. This often means adopting and implementing technologies that will make the SME’s offering more attractive to an OEM.”
A key step in getting your business ready for AI is to finish digitally transforming your processes, advises Professor Tim Vorley, Chair in Entrepreneurship at Sheffield University Management School.
“AI and machine learning are great, but it’s rubbish in, rubbish out,” he says. “If you haven’t got well-curated data, then the ability to maximise AI and machine learning more broadly is really quite constrained.”
If you want to use AI, you need to make your business data digital, consider moving IT systems into the cloud and think about where IoT sensors will be useful.
On a sector level, it’s about embracing data sharing. Sharing information and insights across the UK manufacturing sector (while observing privacy regulations of course) will build data trusts – essentially, archives of anonymised information – that each individual firm can use to help drive its own insights.
Small businesses tend to be quite fragmented from each other and data sharing doesn’t come naturally, Vorley says, but building these data trusts represents a huge opportunity to increase productivity for everyone.
You’re probably used to the idea of fearing that the robots are coming – and they’re coming for all the jobs. But as one level of jobs disappears with automation, jobs that are complementary to the AI revolution will need to be filled.
“If we haven't got the right training, retraining and upskilling in place, then we're not in a place to harness and leverage AI,” says Vorley. “I think that at an organisational or a firm level, finding people with these skills is the first thing and then deploying them in the organisation.”
These are not easy challenges to overcome, but there is movement in the right direction. Data sharing schemes for SMEs are already popping up in the financial world, while the government has launched a nationwide programme of industry-funded AI Masters courses coupled with work-based placements as a major milestone of the Industrial Strategy’s £110 million AI Sector Deal.
Supply chain management shows how useful AI is in predicting the future. But there is another area where the industrial IoT can tell you what to expect – predictive maintenance.
LG in Korea, for example, already uses Azure Machine Learning to detect and predict faults in their machines before any issues crop up.
Your business can use AI-enabled sensors on manufacturing equipment to predict when the equipment needs to be looked at. These sensors collect data on energy consumption and the how long the machine takes to do a job, compare that to maintenance cycles and scan for possible faults. All that information together can allow you to figure out when a machine needs maintenance or is about to fail. And that insight can not only ensure you’re getting the best out of your machines, it can also reduce maintenance costs.
According to Deloitte’s Digital Disruption Index 2019 44% of organisations in the UK have already invested in AI and by the end of 2020, 81% expect to. While investment in any new technology costs money, with AI, the benefits outweigh the money upfront. And AI is actually often a much cheaper option than you might think.
RPA can be offered as a feature of an existing SaaS package. But it’s not the only “as-a-service” on offer. Robotics-as-a-service (RaaS), which combines AI, cloud computing and shared services, is being touted as the business model of the future and even supply chain management is being offered as-a-service these days. All of these approaches are much more cost-effective than buying your own AI systems.
The numbers speak for themselves. Intelligent automation could up your output by 20%. Using AI to manage supply chains can lower operational costs by 30% and reduce lost sales by up to a whopping 75%.
Smart SMEs will adopt and adapt to what they can use right now, in preparation for when the more hyped tools reduce in price – as they always do.
To find out about other emerging technologies that can help your business, including AR and specialised software - head over to our blog.