From single siloed PLM to Knowledge Graphs and AI

From single siloed PLM to Knowledge Graphs and AI

AI is a new trend. It brings a new wave of technological disruption across industries. Think about Web and Mobile past trends. AI will bring huge dividends for organizations that will be able to harness new technology and introduce AI to their products and business processes.

Everyone is coming after AI and Enterprises are investing in AI.

80% of enterprises have active AI in production today, with Asia-Pacific leading all regions. 48% of Asia-Pacific (APAC) enterprises interviewed report AI is in significant use within their companies, with AI deployed operationally today. North American (39%) and European (31%) also are relying on AI use across their businesses with deployments across their operations. Globally 42% of enterprises seeing lots of room for further implementation and process integration.

EngineeringPLM article – PLM Tools with Artificial Intelligence? already two years ago predicted AI intelligence coming to PLM tools.

I’ve been following different AI applications developed for the last few years. Check some of my previous articles such as AI opportunities for Product Lifecycle Management AI-zation of CAD and PLM and Will AI and Machine Learning change PLM value proposition.

We are still awaiting the AI surge in PLM. Vendors today are pushing clients for IOT, Digital Manufacturing, etc. But at the end of the day all these are not possible without “Intelligent PLM systems”. An Intelligent PLM system should ideally do the following without or with very less human intervention—

Intelligent Decision Making – The system should make data-driven decisions. These tools should make use of artificial intelligence and predictive analytics techniques to gauze future risks and take appropriate decisions. Intelligence Data Processing – PLM tools today are simply Garbage-In-Garbage-Out. There is very little validation of on the data that is entered in the tools today. Also many organizations today do not have a central data lake for concrete data policy. Natural Language Processing and Text Mining – Today a lot of useful data is in an unstructured format. There is a growing need to leverage this unstructured data for better product digitization. Intelligent and Flexible Integration – Today most of the Integrations PLM systems have are hard-wired. There is a growing need to make this integration more dynamic to allow more flexibility in how two systems talk to each other.

PLM tools are gearing up for embracing the AI wave. Manufacturing Industry can be a sweet spot for AI vendors right now. These companies are full of legacy data and tools that can be leverage to generate insights.

Don’t be excited too much with these articles. There are some bad news as well. The Verge article –  Forty percent of ‘AI startups’ in Europe don’t actually use AI, claims report. Misclassification is not a responsibility of startups, but it brings a big question on what is considered as AI and what is needed to achieve the desired AI outcome. Check The State of AI 2019 report () – you can find many interesting things, historic and modern perspective.

MMC’s report found that when companies do deploy artificial intelligence and machine learning, the use-cases are often quite banal. Some of the most popular ways the startups surveyed used AI included chatbots (26 percent of companies) and fraud detection (21 percent). In both cases, it’s tricky to judge exactly how much this technology benefits customers. Chatbots are often annoying to navigate and can be a way of just removing the cost of human customer support. And while fraud detection is certainly useful to customers and businesses alike, it’s more of an auxiliary service than a central selling point.

But as we’ve learned with the proliferation of gadgets like artificial intelligence-powered toothbrushes, just because something says it’s AI, doesn’t mean it actually is or that it even matters. Artificial intelligence is a term that encompasses a wide range of technologies and, apparently, a great deal of hype.

AI can transform PLM, but how?

If we try to reduce AI hype, knowledge graph is foundational technology to transform information into reusable knowlegde. It allows people and machines to better recognize connections and data. A typical PLM database can store and retrieve the data. A famous single source of truth database is a simlpe storage of information in the company. There is rarely something that connects it to “product knowledge”. Where is a difference come. It comes when you connect multiple data sets together. Think about multi-tenant database connecting product information. The more data your store, the bigger is context. Data can be anonymize, analyzed and used for decision making.

Because of the structure, knowledge graph captures the structure of information and its dependencies on people,  processes, applications, data, as-built products and many other things. You can think about Knowledge Graph as a next version of operational data lake and BI.

Context requires connecting information and graphs are complex system of information. Many global systems are using these technologies today to build network oriented business.

So, where is the opportunity for future of intelligent PLM systems? In my view, it won’t come from single siloed systems. The opportunity to come with AI and inelligent solutions will come from large scale data systems and model-driven intelligence. The key role in these intelligent systems will play large scale information networks capturing information about products, their detailed information, structure, usage, configurations and customers. Manufacturing Knowledge Graph will provide a foundation for PLM intelligence.

What is my conclusion? Data is a key element in the process of building manufacturing and product development intelligence. Existing PLM system is single siloed and also limited in the way knowledge can be discovered and generated. The opportunity of knowledge graphs in manufacturing is to organize a network (multi-tenant) manufacturing system to support product development and manufacturing model, connect information about products, vendors, contractors, suppliers and connect it together. Just my thoughts…

Best, Oleg

Disclaimer: I’m co-founder and CEO of OpenBOM developing cloud-based bill of materials and inventory management tool for manufacturing companies, hardware startups, and supply chain. My opinion can be unintentionally biased.


Share This Post