Manufacturing has been evolving smoothly since the industrial revolution experiencing some leaps when first electrification was introduced, then computarization of tools and more recently computarization of the whole plant (RPA Robotic Process Automation, still under deployment). Additionally, we have seen leaps in processes first with the assembly line then with lean manufacturing and zero warehousing, A third component in the evolution has been the adoption of new materials (plastic, alloys, carbon fibre, …).
We are now entering a phase of evolution where data analytics (leveraging IoT), artificial intelligence (in tools, product and in processes), additive manufacturing, alloy design and industry 4.0 (with continuous feedback and feedforward across the value chain including the end user) are converging. Add to this the distribution of knowledge across humans and machines and you get a perfect storm.
I’ll pick up the “knowledge sharing across human and machines” in a later post. Let’s focus on the others and consider the macro implications.
The interactions across the value chain influence the way manufacturing is done and “designed”. Whilst manufacturing has been basically a closed system with well defined interfaces with the supply chain and delivery chain, in the future it will have to take into account signals arriving from the products as they are being used as well as their evolution. In fact, products are delivering functionality more and more through software (both embedded in the product, in the cloud and … in the future in their ambient -embedded in other products). This will both provide more flexibility and increase complexity (also because of the partial loss of control).
We are already seeing sign of this happening: more and more airplanes are no longer equipped with screens to provide entertainment. Airlines assume, rightly so, that everybody has a personal screen (tablet, smartphone) hence they only provide connectivity. Home appliances assume you have WiFi and can be programmed/used through apps that are on your smartphone (just bought a washing machine and discovered that I needed a WiFi to operate it!). This is a trend that will grow stronger in the future, both in terms of being used by more and more product and in terms of delivering increased functionalities that will evolve over time.
Data analytics has been part of the manufacturing process for several years (e.g. in the area of quality control). In the coming years it will extend to include data from product operation and those data will be leveraged to provide additional services like:
- fine tuning of performance as result of global data analytics
- fine tune of performance through customisation taking into account that specific context and way of using the product
- release of updates to fix bugs
- release of (at a price) increased functionality
- customer care through the product (you no longer call a call-center, it is the product that connects you to the specific needed support).
- marketing tool (the product suggests new products to buy)
Artificial intelligence is already part of the manufacturing processes (design, production, supply chain management) and of the products/services. It will keep expanding in role and impact in the coming years. In manufacturing AI will impact design (including new alloys search and creation), resource allocation, predictive maintenance of the assembly lines, process re-engineering. Digital Twins (of robots and of products) are rapidly evolving to stage 4 (and 5) and will be embedding AI, with a rise of Cognitive Digital Twins.
AI low-cost chips will become part of products (here the low cost is crucial!). At the same time the increased flexibility and variety of behaviours supported by AI will challenge customer support (what is going on?!). It will also have regulatory implications (privacy of data, accountability…). This is an area where competitive aspects are tackled differently in different geo-political areas. As an example the recent AI strategy adopted by the UK Government focuses more on competitiveness than on privacy protection whilst the EU approach is more focussed on privacy rights.
The biz relevance of data and AI is fostering the adoption of “industry clouds” where data, their analytics and AI can be supported based on that company policy -and biz drive.
The crucial role of communications required by RPA foster the adoption of private 5G in the plant. Furthermore, the servitization of products assumes the availability of a pervasive and effective communication fabric (5G is nice but not mandatory). What would be good is the availability of 5G advanced architecture supporting session control in the devices (products) since this will provide increased flexibility in the orchestration of an ambient (cooperation among different products). This, and the extended use of AI, will also require the adoption of common data ontologies, of shared data spaces, something that is fostered by Gaia-X.
All these transformations require different set of skills and a different market approach: in other words manufacturing has to re-invent itself to both leverage and survive the wave of change.