Part IV of a four-part series.
The machines were stupid. They didn’t make decisions and couldn’t learn anything new. It changes quickly.
With the integration of software and sensors, a machine can come into contact with its environment, and human “intelligence” can be introduced in the form of decision support algorithms. This allows the machine to react to its environment and make decisions.
A very simple pre-computer example is the loom that formed the basis of the future Toyota company: the machine could detect when a tread had broken and would stop the weaving process. Simple but smart: the machine detecting a problem and stopping the process, it no longer needs to be constantly supervised by a human.
Modern cars can distinguish the front from the back of other cars and determine their location, then calculate whether the other car is on a collision course or not. Based on built-in decision rules, the car chooses whether to warn the driver, take control and slightly modify the route or even make an emergency stop. Clever. But never smarter than the intelligence a human put into the system.
Sensory awareness helps make smart decisions
Intelligent machines have sensory awareness of their surroundings and can make decisions based on decision rules built into the equipment. The more it works, the smarter the machine. But the system itself does not learn; people do this for the system.
Machine learning. When an intelligent machine is able to modify decision rules to improve its decisions, it is called “machine learning”. The machine is able to learn and become better and better at its decisions.
Suppose the car in our example modifies interventions based on what it has learned from previous interventions: the machine “learns” from previous experiences and gets better and better.
Machine learning is a great way to not only quickly improve machine response, but also make it responsive to changing conditions.
Although the learning algorithm remains the same, the decision process may become more accurate and adequate, but the machine will never outperform what the learning algorithm allows.
Artificial intelligence. It becomes exciting when the machine is also able to enhance the learning process itself. The machine could start looking for patterns on its own and look for other parameters that influence it. For example, he might discover that a sudden change in rear tire pressure (heavy load) affects braking and steering behavior and learn that “when that happens, I’d better intervene sooner and faster to keep the car under control”.
In machine learning, “learning rules” are static: “If this, then that”. In real intelligence, however, the rules of learning consist of principles. The machine applies these principles to its reality and learns from the result of its response what would have been the best decision, then formulates new decision rules based on that.
These self-learning systems are getting better and better, but for now we still need smart people. And as the 737 Max debacle taught us, intelligent systems are not a solution to fundamental system flaws.
Are “smart” and “intelligent” always a plus?
It’s like with people: there are super intelligent people who do almost nothing with their abilities (they are not very intelligent) but there are also people with little intelligence, but who are intelligent in the in the sense that they optimally convert the little they have into in order to become better.
Ideally, the machine has enough intelligence on board to improve its decision rules and turn those decision rules into beneficial decisions, which then lead to better performance.
1 . The labels “smart”, “intelligent” and “4.0” are unfortunately buzzwords wrongly placed on all kinds of products. And while the technology can be very helpful, don’t expect it to miraculously solve problems that should have been solved without the added complexity.
2. Adding investment and complexity to solving problems is rarely a good idea; simplifying, standardizing, stabilizing, engaging people and reducing non-value added activities are good ideas.
3. Adding new technology can create value when it helps create or achieve goals such as accuracy or speed that could not have been achieved without that technology.
4. As long as machines are at best somewhat intelligent, the people around the machine must be able to convert intelligence into intelligent decisions. And when such decision rules cannot be integrated into the machine, they must be integrated into the process: This is the “A” of the PDCA (Plan Do Check Act): Standardizing an improvement that results from continuous improvement.
What to do now?
In each plant, choices must be made and priorities set. By applying further process improvements leading to empowerment – automation with a human touch – and/or simplification, revenue can be generated without capital investments, while laying a solid foundation for new technologies to come,
First, deploy (new) technology only to create (more) value; do things that bring more value to the customer, and which could not have been done without this technology. Never use technology to organize issues; instead, work to eliminate them.
Visualize losses (ex. everything that your customer is not willing to pay) and address their sources and root causes. Quit “palliative measures” and “organization” of structural problems.
Get your processes in order first; make them fast, flexible, reliable and efficient using the know-how and creativity of your crew. Only then will you start automating for quantitative or “additional product functionality” reasons. Automation is not a solution to bad processes!
Don’t make manufacturing a domain for engineers, management or, even worse, suppliers. Empower everything employees working to make inefficient processes work. They suffer the most from these processes and have the most inside knowledge about them, and are therefore in the best position to come up with ideas to change them.
Put technology at the service of people and processes. Let technology create value! As fascinating and even tempting as some new technologies may be, ultimately the question is: do they really help to achieve fundamental sustainable improvement, reduce risk and make the organization strong and ready to meet our challenges? Is the new technology fixing a problem or is it really adding value?
Keep things on a human scale: if you can’t explain it to your neighbor, it’s probably too complex.
Arno Koch has over 25 years of experience in process improvement and control. Its improvement goals are defined in terms of “halving” and “doubling”. He teaches process improvement at the CETPM of a German university, is a partner at OEE Coach BV and owner of Makigami BV, and has written three books on OEE and two on Monozukuri (“the art of making things”) ).