Having a lot of data is good. Knowing how to use it is better! This was the central theme of the final episode of the L-DIH Talks series held in early June.
There is a proverb that says “plenty is no plague”. However, having plenty of data is not necessarily a good thing when you don’t know how to use and exploit it… not to mention simply extracting and making it available.
30% higher returns
“The explosion in the amount of data and the limits of the capabilities of traditional tools such as spreadsheets to manage it are impediments of progress”, explained Yannick Bruck, CTO of the IT consulting and services firm Fujitsu Luxembourg, during L-DIH Talk #9. “Added to this is the problem of lack of compliance and of available skills. This also makes it difficult to monetise this access to knowledge, as it is sometimes difficult to prove the value of the investments that need to be made. At the end, you realise that data is a strategic asset, but difficult to tap into.”
This probably explains why 99% of companies say they are trying to become insights-driven, but 30% actually report succeeding.
“Data management in silos is another problem,” says Mr Bruck. “This creates additional costs through redundant or excessive processing and leads to fragmentation of data governance and compliance management.”
The solution is therefore based on a single and unified platform, allowing the entire processing chain to be tracked, from the capture of the data to the implementation of process based on these data, through the construction of predictive or descriptive models.
“A cohesive data strategy can generate up to 30% higher returns,” he said. “But today, 81% of companies do not understand what data is required to develop artificial intelligence processes, for example, and 80% of this data is either inaccessible, untrusted or unanalysed.”
For his part, Philipp Coenen, Senior Project Manager at Stremler AG, a consultancy firm specialising in optimising corporate value chains, is aware that “processing data or managing it correctly is a huge competitive advantage today”.
However, he also knows how difficult the task is. “Usually, this data is more or less structured according to the needs and what it is intended for. Often it is slow and tedious to process, and blocks a lot of resources for other operational tasks.”
The idea is, above all, to ensure the availability of the most relevant data, which will allow for linking together the different elements of the value chain. At this stage, therefore, the important thing is not completeness, but quality. “Production processes are driven by customer needs,” he said. “It is therefore essential to have data around three main pillars: sales and distribution, production planning and material management.”
To carry out this approach, Mr Coenen recommends the implementation of a digital twin: a virtual identical reproduction of the entire process – from the initial order to final delivery – making it possible to envisage real-time simulations and to measure the impact of changes in one or more initial data (concerning the products, the quantities to be produced, delivery dates, etc.).
“Digital twins puts the management in the proactive control seat.”
AI for dummies
At this stage, the implementation of processes using artificial intelligence (AI), for which data is the primary feeding element, also come to mind. In this field, certain prejudices are still very much present. “Artificial intelligence is often perceived as complex, unpredictable and difficult to implement. This is not necessarily the case,” explained Benjamin Hourte, Technology Director at EarthLab, a company developing solutions combining big data technologies and remote sensing. “AI requires the application of defined procedures and the use of tools that guarantee data quality. It can be easily deployed using recent innovations in terms of public, private and hybrid cloud solutions”.
Two major axes underpin the development of AI: the optimisation of existing procedures, with a more flexible approach, and the opportunity to explore new horizons and new uses for the data collected and processed. “The success of AI will come in the ability to make it simple to experiment and implement such approaches,” says Mr Hourte.
The fields of application are very broad: quality control, forecasting breakdowns, predictive maintenance of equipment, forecasting production demands, detecting fraud, etc. However, there are some pitfalls to be avoided. “One of the main reasons for the failure of such projects is the quality of the data processed and the difference between the initial wishes and the final reality. It is also important to be careful, first and foremost, that the purpose of AI is well defined: asking the right questions at the outset will increase the chances of success.”
Do not neglect the human aspect
As in every episode, the 9th L-DIH Talk ended with a testimonial from an industrial company that has implemented automated processes. Sensor manufacturer IEE, with its world headquarters in Bissen, in the north of the country, presented its entire strategy for digitalising its production processes.
“Before this automation, data extraction was tedious and time-consuming and we had a reporting rhythm that was generally monthly, which is insufficient,” explained Fabian Leponce, Process Development & Optmization Manager at the firm. “Now, access to data is daily and the creation of reports is almost instantaneous and everything is sent directly by email to the teams concerned.”
This concerns not only internal production processes, but also customer relations, particularly for complaint management. “Each request is documented in our ERP system and the processing of files is much faster than before. This also allows customers to see how seriously their requests are being handled.”
The process of automation and digitisation is very simple, even obvious on paper, but does not start from scratch, the most crucial part being the identification of the right indicators that will be used and monitored over time, and then their extraction. “And the human aspect should not be neglected. It is important to demystify the process and show the importance and advantage of choosing and using the right data. This clearly involves the collaboration of everyone.”
See you in September
The L-DIH Talk #9 was the final webinar of this first season of events. To close it out, a face-to-face event in September will be held to take stock of these four months of presentations and testimonials.
It will also be the occasion for the launch of Season 2, on the theme “Digitalisation and Sustainability in Industry 4.0”. “We are going to look at the best ways to combine digital transformation and sustainable development in manufacturing companies, production sites, etc.,” pointed out Marina Guérin-Jabbour, the manager of the Luxembourg Digital Innovation Hub, who initiated these conference cycles.
Replay the Webinar here!
Back to News List