AI has become common, but will industry follow?

Industrial gears.

While the artificial intelligence hype is plateauing, AI has become a common part of consumer products. Search engines learn from every search query, cameras automatically optimize images, robot vacuum cleaners assist in cleaning our homes and cars provide driver assistance. While AI has gradually become a significant part of our everyday life through consumer products, it’s important to note that they are a result of long-term, continuously ongoing, product development initiatives. Will the same happen also in the industrial sector?

By example, machine learning (ML) algorithms won’t alone make search engines intelligent. The key driver behind progress has been a much broader change, digitalization. An advanced search product that offers the most relevant information in the fraction of a second isn’t the result of just an AI algorithm (sic). The decisive part of building a search product has been to place algorithms at the core of a digital product. Human-generated text and browsing patterns are used to train algorithms, and scalable IT infrastructure enables serving search queries with low latency. Digitalization, scale, and algorithms have together made the search engine an intelligent product.

When this premise is contrasted to e.g. the manufacturing industry, it’s clear that the widespread adoption of AI will require an even more substantial wave of digitalization. In essence, the challenge boils down to which tools and platforms will facilitate developing applications and processes to utilize the very heterogeneous data that has been generated at the production line as well as the entire supply chain. Data or algorithms don’t create value as standalone components. Even after installing a wide range of sensors to a production line, the data they collect is merely a prerequisite for production optimization. Systematically improving production processes requires data refinement, combining data sources, and connecting outcomes to relevant metrics or feedback. A smart production line should be viewed as a digital product, and its development as a product development initiative.

Industrial AI is full of opportunities

The threshold to get started is low. With the progress of digitalization, industrial companies can increasingly and more easily benefit from AI. In fact, it is already an invaluable tool in certain business-critical processes, in which ML algorithms power features that can’t be programmed with rule-based logic. Most often, AI performs tasks that we humans could also carry out, but tirelessly, more accurately and orders of magnitude faster. We humans are left with the role of verification – we accept or reject the proposals of an AI-powered product. 

Visual quality control is a concrete use case that has helped industries like wood processing reduce both waste and product returns. When AI analyses 2D and 3D video streams, we can ensure that products meet quality standards (e.g. correct dimensions, no cracks or other defects), and faulty units can be automatically removed from the production line using robotic arms. The same ML technology can be similarly used to monitor production equipment condition, e.g. for signs of mechanical wear and tear, which in turn enables predictive maintenance.

Although AI is in these examples applied to a small part of production processes, its effects on the company’s value chain are extensive. By identifying and correcting problems as early as possible, significant resource savings can be achieved both in logistics and supply chains. And as AI continuously learns from new data, this implies that early adopters will gain an ever increasing competitive edge.

Present and the future

AI will without a doubt impact the entire industrial value chain, from raw materials to product development, and from logistics to production. The factory of the future is intelligent and only rarely makes a mistake. Little waste is produced, repairs are replaced with maintenance, and each watt is used efficiently. Digitalization is the key that will unlock the full potential of AI. But success will require long-term investments into the development of digital products. And these are also the same products in which the AI algorithms will eventually be incorporated.

Interested in learning more about industrial AI? Read about Silo AI's collaboration with Betolar.

This article builds on Silo AI CEO Peter Sarlin's op-ed in Finland's largest business newspaper Kauppalehti.


No items found.

Want to discuss how Silo AI could help your organization?

Get in touch with our AI experts.
Peter Sarlin, PhD
CEO & Co-Founder
+358 40 572 7670
Peter Sarlin, PhD
CEO & Co-founder
Silo AI

Peter Sarlin is the CEO and co-founder of Silo AI, one of Europe’s largest private AI labs, and a Professor of Practice in applied ML and AI at Aalto University. He has spent his career at the intersection of academia and industry, deploying state-of-the-art AI into products of large corporations and startups. Peter has a PhD in applied machine learning and a pedigree as research professor/associate from top universities like Imperial College London, London School of Economics, Stockholm University, IWH Halle, University of Technology Sydney and the University of Cape Town, and has previously worked for the ECB and IMF, among others.

Share on Social
Subscribe to our newsletter

Join the 5000+ subscribers who read the Silo AI monthly newsletter to be among the first to hear about the latest insights, articles, podcast episodes, webinars, and more.

By submitting this form you agree to the processing of your personal data by Silo AI as described in the Privacy Policy.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

What to read next

Ready to level up your AI capabilities?

Succeeding in AI requires a commitment to long-term product development. Let’s start today.