header_corporate-blog_2019_1920x559px

How we can learn from data

teaser_e-mail-newsletter_ana-campos_3

The ability to use data intelligently helps to tip the scale more and more often. What do companies need to be able to do this? We sat down for a chat with our Co-CEO Ana Campos about the opportunities and boundaries of our data-driven society – and about the fact that data can sometimes be lifesaving.


“Using data intelligently” sounds great, but what exactly does that mean?

It basically means nothing other than the ability to draw findings from data that can be used to make decisions for action. Data alone are not very useful. The key is to connect them cleverly and analyse them. Only then do they offer value and help to make processes more efficient or to recognize risks and opportunities early on.  

What does it take for companies to be able to use their data intelligently?

Companies should know what they could be using their data for. Identifying this potential is the first major step. Possible goals can then be derived from this. That might seem pretty trivial at first, but it really isn’t. Companies are often overly hasty to find a technological solution without being clear on what the goals of the business even are or could be. It’s also important for companies to know their data pools and sources well and potentially tap these. It should be clear, for example, whether the data need to be protected and in what legal scope they can be stored and processed. And finally, companies should also be willing to process data in the cloud so that they are able to benefit from the opportunities of new technologies such as artificial intelligence.

What exactly does the intelligent use of data look like?

For a tunnel-building company, we developed a solution that automatically detects and classifies cracks in tunnel walls – a process that would usually costs tunnel engineers a lot of time and concentration. We used a machine learning model built around Azure Custom Vision that we tested, trained and optimized using real and false cracks in tunnel walls. Metadata for the images can also be used to identify the crack’s exact geolocation.

Another example is the smart bus depot solution we developed for a transportation company in western Switzerland. It uses artificial intelligence to automatically send the buses to the optimal position in the bus depot. This way they are parked in exactly the right order in the evening as they will leave again in the morning, which makes the buses more punctual overall. They can also be used more flexibly as needed by the customer.

Trivadis supports not only companies with the intelligent use of their data but also non-profit organizations such as the children’s home AtemReich. What kind of solution was used there?

The children at AtemReich receive artificial respiration for a number of medical reasons and require intensive care. Many of them are not able to speak, and for some, communication is even more limited. The care workers had previously only gotten indications of how the children were doing from their vital signs. These were recorded manually from the monitors onto paper and were not always connected. Analyses were only possible to a very limited extent due to the vast amounts of data. Which is why we developed a cloud-based solution together with care workers, doctors and medical device manufacturers that connects the nursing records and the medical devices as well as the data from these and makes them available for long-term analyses. An AI algorithm will also make it possible to analyse the data and recognize irregularities in the vital signs early on.

How exactly have the children benefited from this? Can you give us an example?

We were already able to help the 13-year-old Maxi, for example, with this solution. He often had aggressive phases in which he would hurt himself and we were unable to get through to him. We had no idea what was causing it. An analysis of his vital signs showed that Maxi’s aggressive phases might have been caused by him getting a dose of his medication that was too high. The increase in the dosage was so low that we did not even consider it as a possible cause at first – until we saw the connection black on white in front of us. Since the medication was adjusted, Maxi has been doing better.

Are there also examples of the intelligent use of data at Trivadis itself?

One thing we did was to implement a smart Skill Supply Chain that lists all of our employees and their individual strengths. If we need an Azure architect in Munich, for example, then we can use the Skill Supply Chain to find out in just a few clicks who would be a good candidate. The Skill Supply Chain also shows us how busy our employees currently are, which makes it easier to schedule projects.

Another example is the COVID-19 Dashboard, which we developed in March. It integrates and visualizes all the key figures, including guidelines from the authorities on COVID-19. We also made this dashboard available to the public so that other companies can benefit from it as well: www.covid19-dashboard.ch

This might be a somewhat philosophical question: Where are the boundaries of a data-driven society?

The current COVID-19 situation has shown that data and the findings derived from them can help to overcome (extraordinary) situations – and that we are willing to share data about ourselves if it helps the greater good. But it has also shown that data and technology are not everything – that it comes down to the person using that data. And that humans can never be replaced.

And finally, a personal question: What solution could you personally really use?

Many parents are probably in a similar situation as I am: I have two boys aged two and four, so I would love to have an app that uses smartphone photos to detect automatically which Playmobil castle or ship this door belongs to. Right now, we have to come up with creative solutions – but maybe that’s a good thing, too. (laughs)

Are you also looking for a way to use your data intelligently? We can develop the right solution for you together.

campaign_missing-piece_business-automation_header_unterseite_en