How a ‘materials lag’ slows product innovation – and how AI can solve it

Greener, cleaner and dramatically higher performance manufacturing is the promise of Industry 4.0.

McKinsey & Co. defines Industry 4.0 as the next phase in the digitization of the manufacturing centre, driven by four mega-trends: the rise in data volumes, computational power and connectivity, machine learning and analytics, new forms of human-machine interaction, and improvements in connecting the digital and the physical, such as in 3-D printing.

While most people focus on products, materials are the silent enabler of the features and functionality that we all respect in the most popular and advanced products of our time.

Take Apple, for example.

When Apple launched the Apple Watch in 2015, touting its many features and how it would revolutionize the user experience of a mobile device, the company released three advertisements, entitled Aluminum, Steel, and Gold.

Materials centre stage

These ads emphasized the materials they selected over any single feature.

As if to drive home the point further, the following year the company released another ad with the next generation of the watch entitled Ceramic.

The materials that Apple chooses for its flagship products are core to the definition of the product, and, moreover, Apple’s prowess in pushing the boundaries of the manipulation of materials is a key competitive advantage that allows the company to charge a premium.

Apple is not alone.

Boeing differentiated its 787 Dreamliner by constructing the body out of a greater percent of carbon fibre composite than any aircraft before it.

That material allowed for greater fuel efficiency and a better flyer experience, making it the fastest selling wide-body aircraft in history.

At the same time, Ford worked with Alcoa to develop the material–joining techniques to enable the construction of a lightweight, high fuel economy F150.

Speed of development

Companies now are often defined by how fast they can develop the materials that enable their customers’ next generation products.

Corning’s Gorilla Glass is a dominant player in high performance mobile phone screens: making that product line alone worth over $1 billion per year.

Yet, such a huge hit comes with its risks.

Corning faces constant pressure to come up with ever more cutting edge inputs for consumer electronics companies to design into their products. The company only succeeds by its ability to move faster than its competitors.

While this might seem like a standard competitive dynamic, there is something different here.

Ideally, Corning could develop a new glass product each year for each smartphone generation.

However, taking a new material from conception to an actual product takes 15-20 years.

Materials lag

Today, materials development cannot be so far behind the product development cycle.

This materials lag creates huge risk for these companies and lost opportunities across every high-value manufacturing industry.

The way to erase the materials lag lies in the ability to analyze data from various sources using machine learning and analytics.

While all four mega-trends apply here, two stand out as the enabler of materials design.

Data

Materials data is often found in various sources, scattered far and wide, and lives in various unconnected formats.

The first step in solving this problem is to organize the world’s data as it relates to materials and chemicals.

The days of having a product developer go and pull a book and figure out how to use that data are gone.

The industry requires a materials Rosetta Stone to connect the disparate sources of understanding.

Bringing these diverse data sets together opens the door for powerful analytical tools to drive the materials lag to zero.

Machine Learning and analytics

Even with all of the world’s materials and chemicals data in a single place, it is not possible to use naive Artificial Intelligence (AI) to drive practical materials development: the data is simply too small.

Whereas Google and Facebook rely on millions of datapoints per second to inform their artificial intelligence, most materials companies generate measurements on hundreds or thousands of materials per year.

As such, the best materials AI is optimized in order to take advantage of known relationships in chemistry and physics, and to combine these with computational simulation, traditional analysis, and experimental data to form a powerful prediction engine.

Ideal assistant

This essentially gives a computer the training of the ideal lab assistant – an understanding of the lessons we have learned in materials over thousands of years of development, but with a north star of high-quality data to connect that theory to reality.

This type of engine can be used to predict the performance of known materials under new conditions, optimize new materials for known applications, or optimize the development process to reduce the risk and total number of iterations to achieve a product goal.

This will eliminate the materials lag that slows down next generation product opportunity.

Materials innovation is the key enabler of the next generation of high efficiency products from energy, to aerospace, to defence, to safety and comfort.

But the materials lag in product development threatens to dramatically slow that innovation.

By bringing the data and analysis advances that power Industry 4.0 to bear on the materials development process, artificial intelligence promises to propel the next generation of products, just as Industry 4.0 is propelling the next generation of manufacturing.

For more information about the 2017 class of Technology Pioneers, visit our website.