The real world and its digital twin are collaborating to bring forth something called “mass customisation”, a new manufacturing culture which, as the term suggests, will be the basis for the most diverse ecosystem of engineered products ever seen.
To understand and articulate some specifics about what’s being called “the new frontier of manufacturing”, one of the world’s largest management consultancy firms, Deloitte, partnered with the Singularity University, a forum for technology futurists co-founded by Ray Kurzweil, to organise a conference called Exponential Manufacturing, featuring many thought leaders working in the industrial sector.
The Singularity University says the idea of the event was to bring together the world’s top experts in technology and manufacturing industries to help participants gain a deeper understanding of the core technologies reshaping manufacturing’s future, including:
- artificial intelligence;
- robotics and unmanned aerial vehicles;
- big data;
- synthetic biology;
- digital fabrication;
- smart sensors and networks;
- and machine learning.
While each item on the above list is a huge subject in and of itself, experts in the field of manufacturing and technology who took part in the Exponential Manufacturing seminar say they can all be summarised in certain ways, as explained in a video produced for the event (above).
Peter Diamandis, co-founder and chairman of Singularity University, says: “The ability to go from intentionality – what’s in your mind, what’s in the consumer’s mind – to actually having it right then and there, fast and cheap, is ultimately what all these technologies are converging to make happen.”
Often, all these diverse technologies – as they are applied to manufacturing and related sectors – are indirectly referenced in the catch-all term Industry 4.0, which itself refers to the idea that the world is going through the fourth industrial revolution.
Just to recap:
- the first industrial revolution was brought about by mechanisation and steam power;
- the second industrial revolution was essentially about electrification and mass manufacturing;
- the third industrial revolution is said to have been led by computers and automation systems; and
- the fourth industrial revolution is where everything is connected together into what are called “cyber-physical systems” – meaning the integration of software and hardware and even humans.
“Industry 4.0 really references the next revolution in manufacturing, which, simply defined, is mapping the physical and digital world together,” says Steve Shepley, principal at Deloitte, and one of the speakers at the Exponential Manufacturing conference.
Delving into the details
RoboticsAndAutomationNews.com briefly caught up with Shepley for a deeper insight into the subject, and to ask what the priorities are from a manufacturing company’s point of view.
In this exclusive interview, Shepley – whose specialism is manufacturing and who advises many of the world’s largest companies – delves into the details of what he considers to be an exciting topic, and the most effective ways it can be applied in manufacturing.
“The digital topic is an exciting one because it exists in what I call the many-to-many space,” says Shepley. “There’s many problems you can solve with it and many different ways to solve those problems, which is always one of the struggles that our clients have.
“Usually, when it comes down to the big pockets of value, it really falls down on the reality that if you consider a highly engineered product manufacturer – and I’m not talking about your average consumer products company but those that make highly engineered products, whether it’s a front loader that runs out on a mining site or a piece of machinery that sits on a factory floor – those companies, if you look at their profit and loss, primarily, almost everything they do sits in two spaces.
“The first chunk of cost all sits within their production and operations environment. If you break that down, most of it is driven and throttled by their people, their direct labour workforce and their indirect workforce that supports the factory.
“So that’s the first, highest and most applicable place for digital – to really go drive up the productivity of that massive workforce that most of the manufacturers own and manage.
“When you look at the cost structure of a highly engineered product manufacturing company, 70 to 80 per cent of that is in the heart of their operations and what they do from a manufacturing standpoint.”
Shepley mentions energy costs, which while significant, they are a very small percentage of the total cost structure of a company, although he accepts it depends on the type of company. A chemical processing plant would have a much higher energy cost ratio, he says.
“The second biggest pocket of cost when you look at a highly engineered product manufacturing company is all around engineering and their engineering talent.
“That engineering side is where all the innovation happens, it’s where all the new products are developed, it’s where all the ideas flow that eventually get made in that manufacturing environment.
“That is typically your second largest set of individuals who are, oftentimes, far more costly than you want them to be because of their inefficiency, their inability to effectively manage designs, bills of material, products that have been billed, engineering changes that flow out to the floor, knowing which changes are worth their time and which changes are not, and then of course how to drive real innovation – sometimes, I think people get stymied by innovation and it drives value out of their talent.
“So when you look at the digital topic, I think its application into that engineering space has huge opportunities for our clients.”
While the two applications he mentions are what might be called “internal” aspects, the third way in which digital can make a big difference, says Shepley, is by transforming customer relations externally.
“Using digital with your customers is absolutely a priority,” he says. “As a matter of fact, it’s pretty much become table stakes these days.”
By table stakes, Shepley means it’s almost a standard requirement.
“For those manufacturers that are behind, using digital technologies to collaborate with your customers, connect with your customers, know them better, respond to them faster, integrate, provide them information better – that of course provides stickiness around the purchase of products and the sale of aftermarket parts.
“That one I would probably put as an absolute priority. The other two I would say manufacturing is lagging pretty dramatically from a digital standpoint.”
The wisdom of numbers
Companies like Deloitte may have started out as accountancy firms, essentially, but since numbers are the essential measure of value of a client’s business, the inevitable logical progression has seen Deloitte become a much broader consultancy service and advisory practice.
The mainstay of Deloitte’s work is driving big implementations that involve technology, process as well as people improvements which all go towards improving the balance sheet of a business.
Such consultancy and advice is becoming more crucial to big business because the digital landscape can be complex and demanding, but it’s essential to keep up with developments because, as Deloitte points out, 50 per cent of S&P 500 firms are likely to be replaced over the next 10 years.
The Standard & Poor’s 500, or the S&P 500, is an American stock market index based on the market capitalisations of 500 large companies listed on the New York Stock Exchange or Nasdaq.
“If you really step back and look at the data and look at the churn of the S&P 500 – certainly over the past five to seven years – a very key trend emerges,” says Shepley.
“That is, those companies that are growing the fastest and reaching what I would say are major market spaces are those that employ two basic tactics.
“One, they are highly digital. Meaning, the way they have developed their products, their services, and the way they run their organisations are all built around a digital operating model.
“The second piece is that they tend to operate in what’s called an asset-light environment. Meaning, they’re able to capture the benefit of many assets distributed across an ecosystem, but bundling those together using some type of digital means to create that value.
“The quintessential example is Airbnb, being one of the world’s largest providers of short-stay accommodation, or rooms, and yet not owning a single hotel.
“But that’s just one example in one specific space. There are companies today that aggregate latent capacity across lots of manufacturers. By just putting that together and offering it on one single platform, they are able to get quite a premium within their markets. That’s an example of a digitally connected company that’s extremely asset-light.
“So any of those laggards within the industries today that aren’t embracing digital and aren’t understanding how to go more asset-light, or to get more utilisation out of existing assets, those are going to be the losers.”
Automating the digital process
Some might say the digitisation process is the merely the first step towards automation, wherein processes which used to be conducted or managed by humans are now undertaken by robots.
Some might call this robotic process automation, or business process automation. But while those terms may be familiar to some, the specifics about them are not widely understood, unless you’ve had some experience of those automation systems.
Digitisation is relatively easy to understand. A journalist who in the past used a typewriter would produce what was called a “hard copy” of a story – that is, the type-written story on a piece of physical paper. But starting a couple of decades ago, the same muckraker may have had his state-of-the-art typewriter replaced by a computer and only needed to move a “soft copy” of the article from one computer to another within the network.
What is it that’s being automated? Apart from the carriage return, that is.
Shepley tries to explain in the context of manufacturing, although the explanation would be applicable to other industries as well.
“The opportunity and benefit of moving to automation is very significant, certainly in the manufacturing environment,” says Shepley. “As I mentioned, most of the costs within the profit and loss of a manufacturing company is either in their direct labour, their indirect workforce or engineering – so it’s all about the people.
“And if you actually take the time to actually break down how a company works, what you very quickly find is that most knowledge workers within a company spend very little of their time actually creating knowledge and making decisions and taking actions.
“The preponderance of their time is actually spent trying to find data, put data together, or process transactions in enterprise resource planning systems. So, things that are extremely repetitive, highly manual, and very logic-driven are somewhat complex. Meaning, you need the human brain to be able to tie all these various components and pieces together. And they actually tend to spend 70 to 80 per cent of their time doing that, instead of spending 70 to 80 per cent of their time adding value.
“There are all sorts of things that are digital today. The problem is that you need to automate in a closed-loop system from a demand signal all the way to an action happening in a company.
“Where robotic process automation becomes very powerful is its ability to trigger an event from the point at which an email from a customer, or an email from someone internally from your company, for that email to be read by a bot, and for the direction of that email to be decided by a bot using natural language processing.
“By reviewing that email, the bot is able to go and log on to one or many of your business operating systems and actually process transactions that need to happen, and go all the way to the end and close the loop on that action.
“So, automation is really bridging together a lot of things that are already digital, and tying them together as a human would. And what’s powerful about it is the cognitive aspect.
“A lot of the time you think of automation, it’s more of a swivel-chair task, where the same person types something in one system and then swivels their chair and types the same thing in another system. Automation can eliminate such instances – and that’s not a small amount of work for a lot of our companies.
“But they can also do a lot more advanced things, where you can build in your ‘if-then’ logic. You can also put a cognitive engine in there that can evaluate the context of the decision being made, know how risky it is, look at previous decisions that were made and actually process that information as well – all within the cognitive logic within that automation.
“Automation does have some of the capabilities which reside in the human brain, but there has to be some logic, order or structure by which those decisions are made.”
The application of business or robotic process automation are “extremely broad”, says Shepley. While many might think the main applications are in the back office, in finance or accounting, because there are a lot of repetitive, logic-driven tasks required in those departments.
However, in manufacturing, there are many more potential applications for bots, autonomous programs which operate on a network and interact with users and systems. “If you look at operations, everything from the ingestion of a demand signal all the way to the production of a product for your end customer, there is a whole swathe of front-end planning and analysis tasks that can get automated – tasks that today take a large amount of human resources.
“A lot of the complex work done by mechanical engineers, industrial engineers on the factory floor – a lot of that type of work can actually be automated.
“A large portion of engineers’ time is spent on low value-added tasks, simple things, like management of the engineering change process, various review boards, approval of engineering specs – all of those tasks can be automated within the engineering space.
“A lot of your supply chain communities can be automated. Today, you’ve got armies of people working on purchase orders, checking the status of material stocks, managing logistics – all of that can be automated.
“The same applies to IT, HR, so on and so forth.
“The real opportunity for automation is actually in the heart of the business, the operations. We just barely touched and tapped these areas.”
He says bots these days are capable of collating vast amounts of information to provide accurate answers to highly complex questions, and perhaps the biggest two advantages of bots are that they don’t make mistakes and they don’t take breaks.
Learning from the leaders
While all these digital processes are relatively new, a lot of companies have already adopted them. In particular, the auto manufacturers, and what are called Tier 1 suppliers – the component manufacturers who directly supply the top manufacturers whose branded goods are mostly household names.
But as might be expected, the tech companies everyone will have heard of – the “digital natives”, as they might be called – are way ahead in terms of digitisation for one obvious reason: they’re often the ones building these digital solutions.
Interestingly, continuous manufacturing – oil, gas, chemical processing and so on – may actually be further ahead down the road of digitalisation than discrete manufacturing – self-contained products. The reason for this could be quite simple: a typical continuous manufacturing plant would have relatively few human workers and instead use a vast array of sensors and automate much of the monitoring.
There are well-known, ready-made platforms which enable companies to introduce such pervasive digitalisation into their operations, the most well known perhaps are GE Digital’s Predix and Siemens’ MindSphere.
But, according to Shepley, the reality is that, in order to be really successful in digital, you need a whole range of different technologies integrated together, because there is no all-purpose platform that will suit every business.
“If you really step back and look at this digital space, there is a whole swathe of technologies that you need to be successful in digital,” says Shepley, who declines to give his opinion on any of the platforms available, but provides an overview. “Some of those technologies allow you to connect better to the physical world – the front-end sensors, the 3D scanning – all of these things that connect into the digital space.
“The second thing you need is to capture the data, and a lot of different companies have created big data ingestion platforms, basically giant data staging areas to hold hordes of data and then be able to collapse it and condense it down into meaningful form.
“Then once you’ve collected that data, you have to take that data and turn it into information, and that’s where advanced analytics applies. There’s all sorts of what I would say are basic algebra all the way to machine learning, more advanced algorithm development, and of course into artificial intelligence, and that includes video processing, natural language processing – all sorts of new analytical techniques can now be applied to that data to convert it into information.
“But once that information is collected, you also need to deliver it back into the operating environment – you need to get that information to something physical, whether you send a direct command to a machine to do something differently, or a drone, or you give it to a human in some type of action-oriented form.
“So that’s really how success can be defined in digital. If you take that entire loop, there’s only two power points, or value points.
“One is the source data: the ones that own the source data are going to win in the long run.
“And the second-most powerful point in that entire loop I just illustrated is owning the algorithm. Not just knowing math, but the ability to actually build a custom algorithm that can take source data and then convert it into something extremely valuable.”
Shepley says that, over time, the underlying technology is going to get commoditised: the ability to store data, move data, transmit data, and such things are going to be commoditised.
Source data and algorithms won’t, and those who own and control those fundamental things “will be very successful”, says Shepley.