The crisis we are experiencing has been taken by many companies as an opportunity to rethink what makes them different, even questioning their own reason for being future.

But imagining the future is not something you can do by following a recipe. The external point of view, as may be that of a partner to whom a company entrusts advice, is to be taken into account, but cannot be without the completeness of the information and the perspective of the customer.

In the middle of these two positions there is then a path that formally brings together the people called to think and to think: it is what is called co-creation.

The process, explains Christian Leutner, VP & Head of Product Sales Europe of Fujitsu, starts from an evaluation of the hidden opportunities in that data heritage that most companies hold, thanks to artificial intelligence and data science, In practice: a data-driven approach shaped by co-creation.

The framework for data transformation

There are many examples that can highlight this collaborative approach and effective methodologies to generate ideas, implemented through a global ecosystem of technology partners. But that is not enough. According to Leutner, a framework needs to be created with the clear objective of supporting data-driven transformation. Without it, the process would slow down and probably encounter many problems.

Today, companies acquire large volumes of data mainly unstructured, such as text documents, audio and video files, email messages and images, many of which are generated at high speed from different sources in different locations and under multiple managers.

Gartner expects that by 2022 more than 50% of the data generated by companies will be created and processed outside the data center or cloud: a percentage that is expected to rise to 75% by 2025.

Without a systematic framework, data is little more than a cluster of information disconnected from one another. The challenge is to draw value from them, to monetize them, innovative and generating new business and revenue opportunities.

First, however, it is necessary to rationalise and integrate data through the entire company and its value chain. Flexible, agile and efficient foundation are essential for data science and artificial intelligence to produce true value.

Companies and connections

All this sounds good in theory, but experience says that most companies, despite the awareness of the intrinsic value of information assets, in the initial phase fatigue.

On the other hand, it is not easy to identify connections, understand how to manage information wherever they are (on-premises, in the cloud or, more easily, in both environments), protect essential data from loss risk with adequate security and protection measures, and apply artificial intelligence

This is what happened, Leutner exemplifies, in a project launched for an important European provider of pre-payment systems. The customer, leader in e-vouchers, payment instruments for which neither a credit card, bank account nor any other personal data is needed, had asked support in intercepting money laundering attempts at the exact time when they were committed.

Using the Fujitsu Primeflex for Sap Hana solution, it has been able to drastically reduce the loading time of the mass data needed for the purpose. Previously, the activation of 145 million datasets took several days. Completed the deployment of the new analytics platform, this time dropped to 15 minutes.

Discovering the insights in the data independently

With unsupervised artificial intelligence, an even new and developing field is defined, which allows an artificial intelligence system to question data to determine any different and valuable elements.

To better understand what its potential might be, Leutner gives us another example: the Airbus SE aerospace group has organized a global competition, the Airbus AI gym challenge, to find out what the most accurate use of a system of

The idea is that flight engineers connect large quantities of sensors to helicopter prototypes to acquire every nuance of their behavior.

Since all sensor data are considered normal, the objective was to create an artificial intelligence mechanism capable of operating without prior instructions from engineers.

The challenge was to improve early warning signal detection and accurately identify potential problems in the huge amount of data available, particularly by looking at abnormal data.

For the purpose, Fujitsu has developed a way to exploit unsupervised artificial intelligence and detect anomalies in the accelerometer data of the Airbus helicopters not yet certified for flight.

Fujitsu’s solution achieved a 93% accuracy thanks to the DeepTAN unsupervised artificial intelligence model that takes data sequences from multiple sensors and analyzes them over a pre-set period of time, intercepting the abnormal behaviors of the sensors thanks to an

Fujitsu plans to industrialize this solution for unsupervised analysis of time series complementing DeepTAN with end-to-end features, integrated data pipeline and even more advanced algorithms.

• The creation of a data-driven enterprise defines its strategic direction for years to come. It makes sense to choose a partner who can keep the course. Yet, considering the complexity of digital transformation, no single IT organization or technology company can be expert in any discipline • Danilo Rivalta, Ceo of Finix Technology Solutions Data-driven transformation requires an ecosystem in which many specialised organisations work effectively. We accompany companies in their digital transformation processes, adding the value of a consulting approach of experience and vision in accordance with service levels and performance indicators.

Leave a Reply

Your email address will not be published.

You May Also Like