Semantic Technologies: Making Sense of Data

Introduction

Data driven innovation is becoming the key enabler of competitive advantage in this decade of fast changing business dynamics and knowledge-driven economy. According to OECD report (2015), Data-Driven Innovation: Big Data for Growth and Well-Being the global market for big data was evaluated at US$17 Bn. and growing 40% annually.

However, the promising opportunities presented are overshadowed by equally reciprocating challenges, among which is the usability and even more so relevance of the available data. In fact, much of the intelligence is still hidden in either heaps of paper or in the minds of your work force. In addition, digital data often needs to be aggregated and formatted before it can be used. Not to mention tacit information, which lies beyond the literal meaning of data, the meaning of which has to be derived as well. As a result, the efficiency of your innovation directly depends on your data architecture, or lack thereof.

Ready-to-use information necessitates the data to be transparent (ready to read) and intelligent (ready to process). It directly follows that scattered heterogeneous data has to be addressed firsthand to make it ready to read. On the other side, data intelligence is an extra layer that can be added as a service to any data lake, in the form of a technology that binds data to its meaning. This type of knowledge-based data modelling opens the doors to intelligent, intuitive and even predictive applications. Above all, these technologies will allow human expertise to better focus on data consumption rather than data consolidation.

Data Meaning is the key

In today’s ecosystem, digital data coexists with legacy documents. Both data sources are equally relevant to your daily business and products. Considering this, the information system needs to keep pace with these heterogeneous sources of information and become an intermediary between the user and the data to accomplish the technical objectives at stake. In that way, more system would mean less complexity for the end-user. Well, easier said than done!

Whether it be the format, version or accessibility that differs, a published data seldom is exposed, as in made available for others to benefit. In fact, data is as self-centered as its author is an expert in his field! This contributes to the creation of silos between internal organizations and the added complexity to exchange with suppliers. Again, accessing more data should be an enabler and not a source of confusion. As such, exposing data still remains insufficient.

To turn data into knowledge that would ultimately fuel the innovation we’re aiming for, it is critical that the data makes sense to all your core enterprise systems. To achieve this, the systems should be able to break the coupling between data and how/where/by whom it was initially created. For instance, company investment strategy or product IP could be stored in traditional Excel formats. But how would you evaluate its importance? Access to this data should not be skill or person-dependent. The knowledge it contains should be explicit. On the other side tacit data, open to interpretation, will most often undermine the capability of your downstream systems to make use of it.

To turn data into knowledge that would ultimately fuel the innovation we’re aiming for, it is critical that the data makes sense to all your core enterprise systems. To achieve this, the systems should be able to break the coupling between data and how/where/by whom it was initially created. For instance, company investment strategy or product IP could be stored in traditional Excel formats. But how would you evaluate its importance? Access to this data should not be skill or person-dependent. The knowledge it contains should be explicit. On the other side tacit data, open to interpretation, will most often undermine the capability of your downstream systems to make use of it.

Recommendation

This solution probably relies on a flexible data structure that serves all: people, applications and machines at once. Based on an agnostic framework, it leverages heterogeneous data sources and qualities while avoiding endless replication that’s synonymous with out-of-date and partial information. In other words, a continuous digital thread that connects (but does not transfer) data from all corporate systems. Furthermore, it needs to be generic and scalable to any industry, identifying process-centric objects – bringing the business knowledge back at the center of the information system.

This can be achieved by leveraging semantics that acts as the glue between systems and people, where the meaning of the data, its object, is not only obvious but also tangible and manageable as an asset, bringing value to the company. Ultimately, a single and accessible digital self of a company’s asset can be coupled with a backwards data feed loop from the physical facilities, resources, and data-generating products. That’s the definitive target of a digital twin where data is neither digital nor factual, but only relevant.

Semantic-Tech

Summary

Data continues to grow exponentially and is now massively available but organizations are struggling in driving value from it. Enterprise systems like PLM or ERP sounds like a Stone Age solution for most of the experts, but it remains science fiction to a lot of users who still face data access and comprehension issues. Only if one could bind data sources and make it understandable by its resources, be it an engineer, shop floor worker or a collaborative robot, maybe then could we speak of reducing the complexity of our data streams with a smart, adaptive system.

In my view, this is one of the most disruptive trend and will redefine the way currently products are designed, manufactured and serviced. This digital thread can enable creating digital twin or real time digital clone of the product. It is important to understand that this is a marathon and not a sprint. It needs systematic approach and open mindset to adopt new processes and technologies. I am sure, as companies embark this journey, they will start getting benefits like reduced time to market, optimized cost and ultimately resulting into brining innovative products faster to the market with improved customer experience. Also, it would be interesting to see, the role of start-ups and non-traditional players, as companies are trying to digitize everything and connecting the dots.

The digital thread is going to change our future and business models of the future. We should not be surprised if we see different prominent players dominating the market in future than the traditional or familiar names in automotive industry or to that matter in several other sectors.

With Smart Products, Smart Manufacturing and Smart Services, the world would be a different world, today’s science fictions are tomorrow’s realities and with “Digital” focus, the tomorrow is shaping now. It’s no longer tomorrow!


Thomas Bachellerie – Sr. PLM Consultant

Sylvain Blanvillain – Sr. Industrial Engineer
Stéphane Biard – PLM Consultant

Leave a Reply

Your email address will not be published. Required fields are marked *