If I hear Uber referred to as “a glowing example of the new digital age” once more I think I am going to cry. In my opinion, the real success behind Uber was how they surfed the borderline legality of their model. They created a business that deliberately flew in the face of all the taxi regulators and the city permit offices. In short, Uber’s founders were either going to make millions, or go to jail. How they pulled it off was quite honestly brilliant.
Uber, Airbnb and most of the other technology unicorns are all multi-billion dollar businesses that seemingly came from nowhere, boasting few assets and tiny overheads relative to their valuations. The message is clear: disruptive technology, rapid user acquisition and big data remain the biggest assets in the digital economy. Say goodbye to conventional business models and hello to Digital Transformation.
But if you manufacture a complex product, the idea of owning few assets seems both implausible and impractical. Enter Industry 4.0.
Industry 4.0 was first introduced by the German government as an initiative to computerise manufacturing. For those not familiar with the term, let me quickly explain. Industry 1.0 was the Industrial Revolution, 2.0 was Mass Production (e.g. Henry Ford), 3.0 was computerisation plus robotic automation (e.g. Toyota) and 4.0 is “Cyber Physical Systems”, closely linked to the Internet of Things (IoT).
One example of connected IoT is with manufacturing production lines that are aware of their own health, can predict future failures and therefore schedule their own preventative maintenance. Using Machine Learning they can even look for patterns in their operating data to help refine any prediction. The goal is zero downtime for the manufacturer. However, the reality still falls well short of what could be achieved using the eyes of an experienced production-line manager to examine that data.
While it’s not possible to fully simulate the brain and expertise of a production-line manager, it is possible to improve Industry 4.0 significantly using Artificial Intelligence (AI). We can do this by building in an Artificial Narrow Intelligence (ANI) solution which connects to the available data, applying a model of human expertise, or Cognitive Reasoning. This ANI can kick in and apply nuanced human reasoning on top of the connected, predictive analytics. Without this ANI layer, each production line would still need an expert on hand to interpret the data. Moreover the ANI layer can also augment the expert by helping them to identify a broader impact. In our example this would mean the ANI would identify a shift in the production schedule would impact delivery times. Based on this it could recommend a course of action – in this case perhaps an expedited shipment to make up the lost time.
So within the confines of these narrow use cases, AI can be powerful when it has an awareness of the physical world, which it gets from IoT. Industry 4.0 works better when we consider it in the context of new digitally transformed business models. If we can build human expertise and Cognitive Reasoning into all aspects of industry, then efficiency will dramatically increase. Let’s call this Digital AI Transformation 4.0 (DAIT4.0), just for fun.
So what’s practical to achieve today and where do these phenomena converge to create a perfect storm?
Here are a couple of examples:
Let’s start with the outcome that the customer is looking for. Consider for a moment a company that manufactures diggers. Does the customer actually want a digger or do they just want a hole in the ground?
Some customers need to dig a hole so often it’s convenient to own a digger. Others only need it occasionally so prefer to rent. So how can DAIT4.0 help these customers with differing needs?
Well, when a digger is purchased why not give the buyer the option of connecting that digger to a network of customers looking to rent? The plant owner could use the downtime for their asset as a revenue generating opportunity. The manufacturer could also generate a new revenue stream.
But how would you know when the digger is available, or even if is in a fit state to be used? That’s when IoT kicks in. Information from the digger’s telematics are transmitted to the manufacturer and the availability brokered with a rental marketplace. The manufacturer can even use this telematics data to intelligently predict potential failures and offer flexible, cost-effective servicing “just in time”. These are all revenue opportunities that would not be possible without this technology.
In this case, AI has become both a differentiator and a revenue generating opportunity, increasing value for all.
These are not new concepts in themselves, but there have historically been some practical barriers that have inhibited such models being cost effective.
Firstly, if you are offering a rental service where you are taking a small “booking” margin, how do you administrate that without incurring expensive call centre costs? Also, a predictive maintenance model has to be tuned to ensure that the system does not under or over-service, and the implications of getting this wrong could prove exceptionally expensive. It is not an easy feat at the outset, especially relying on emerging machine learning and predictive analytics technologies which need to be trained.
One solution is to apply a layer of human knowledge on top of these predictive data systems. That’s where the cognitive AI element of DAIT4.0 comes in. It is possible to build a human-like AI layer that can behave as an experienced worker would, looking at a predictive maintenance request and making the right judgment. Likewise, it can provide advice to someone renting the asset on how to better achieve their aims.
This sort of AI may sound like a relic of science fiction, but it is quite viable. There are major plant manufactures that have already implemented elements of this technology already. It’s real and it’s here today.
Does a sick patient want a remedy to an illness or would they rather not be sick in the first place?
Let’s consider a manufacturer of stents. If you are not familiar with this medical device, a stent is an implant that is inserted into a patient’s artery to improve blood flow.
What’s interesting about these devices is they are positioned in a major artery and are therefore useful in delivering medication on a gradual basis to certain patients.
Now consider adding a small sensor to the stent and connecting the resulting telemetry to the Internet giving doctors the opportunity to remotely measure the health of each patient.
Now they have a device that can monitor its own function and can also verify that the drug is being delivered as expected. It even has the potential to monitor for new illnesses.
When a problem is detected the system has the potential to interact with the patient and recommend lifestyle changes that can prevent disease. It could even contact a doctor and provide decision support to assist guiding them to the right answer.
So in this example we have IoT, connecting to predictive models, connected to an AI system. To make this a reality, you need DAIT4.0.
Not only are these technologies used today, start-ups are also looking to take this to the next level and put similar sensors in our clothes. Intelligent clothing is already starting to become viable in athletics where hydration, pulse rate and temperature are measured, correlated and used to drive an individual development program.
Who knows, maybe one day your t-shirt will alert you that you need to see a doctor today to prevent a heart attack tomorrow.
Once you start to ask the question “what does the customer really want?” then this DAIT4.0 concept really starts to open up exciting possibilities including new products, services and business models in multiple sectors.
So it’s an exciting time for business and technology. If we get this right and DAIT4.0 becomes part of our future, it could drive a pace of change that could be quite astonishing.
As the UK and Europe face significant macroeconomic change post Brexit, it may be that DAIT4.0 could herald a fundamental positive impact on our economy. Both automation and AI specifically will change the shape of business and the workforce.
Even if it only achieves a fraction of the predictions, the impact will be significant.
The challenge I would pose to anyone reading this article is simple. If you already have an Industry 4.0, Digital Transformation project or perhaps an AI initiative – stop for a moment and consider the implications of a truly joined-up, blended solution.
You may be astounded where it takes you.
Feel free to contact me if you would like to share what you are working on in this space. Matthew.Buskell@rainbird.ai