We Know How to Work with Data from A to Z, Says Stepan Kopriva, CEO Adastra Czech

September 19, 2024

Data is often referred to as the oil of the 21st century. With the help of data, businesses can gain vital insights into their operations and processes can be optimized and made more efficient. Thanks to data, we can even predict future events using artificial intelligence. However, fully leveraging data’s potential requires an available and resilient infrastructure.

For more than 20 years, Adastra has been focused on solutions that support efficient data management. Today, Adastra is one of the top companies in its field on the market. Thanks to many years of knowledge combined with cutting-edge technologies, Adastra provides clients with comprehensive solutions in a variety of areas, including buildingdata platforms in the cloud, data management and data governance, data analysis, artificial intelligence, IoT, and the development of mobile and web applications. In honour of Adastra’s entry as a partner of the international logistics conference SpeedCHAIN International, we asked CEO Štěpán Kopřiva for an interview.

You are a co-founder of Blindspot Solutions, which is part of the Adastra group, and focuses mainly on the use of artificial intelligence. What were your beginnings with artificial intelligence?

I co-founded Blindspot Solutions with Ondřej Vaněk and Michal Pěchouček, but the trajectory began during our studies. I studied artificial intelligence first at the Czech Technical University (ČVUT) in Prague, and then continued at Imperial College in London, where I studied computer science. This included artificial intelligence, but also many other topics such as the management of large systems, whether industrial or economic. And because artificial intelligence, I would say, has the greatest overlap in economics and other business fields of all computer sciences, I was always interested in its application in companies.

After my studies, I worked as a software engineer for a while, but at the same time, I was pursuing a doctorate where I met Ondřej, and we decided that we wanted to apply artificial intelligence to real-world problems. We wanted to transfer what we had learned at university into practice and help companies solve their real problems. That was around 2013, and the application of artificial intelligence in large enterprises was in its very early stages. So, we started working for clients from various industries both in the Czech Republic and abroad. We worked for large international companies, such as ŠKODA AUTO, but also for startups from Europe and America. And wherever artificial intelligence could be applied, we applied it. From the beginning, we knew we wanted to deliver solutions whose results could be clearly measured, because even back then, we believed that the benefits could be significant. The beginnings were not trivial because especially in the Czech Republic at that time, very few people knew what artificial intelligence was and what it could bring to their company. Today, it is different. We have hundreds of projects behind us where we can demonstrate the results.

In the autumn of last year, you transitioned to the position of CEO of Adastra. What goals did you have when you took up this position?

In 2017, Adastra acquired a 50% stake in Blindspot Solutions. The idea was that Adastra had always worked with data, and artificial intelligence can be used precisely on top of that data. In the Czech Republic, Adastra is a leader in providing digitalization services, whether it is data management, artificial intelligence, software development, or consulting. But Adastra is not just a Czech company; it is a company with a global presence. Last year, we changed the way Adastra operates. The Czech part of Adastra aims to provide high-value-added solutions not only to customers in the Czech Republic and Slovakia but also to customers of the Adastra group in other world markets.

By investing in global centers in the areas of cloud, data, and artificial intelligence, establishing global leadership in key industries, and fully integrating global delivery capabilities into one operating unit, customers anywhere in the world will have access to the services of the entire group. Each of our branches has unique talents with specific competencies. Our goal is to offer customers the best of the entire group.

This transformation will enable the company to continue growing in key markets in Canada, Central Europe, and Germany. It will also support faster market penetration in the USA and Western Europe and the expansion of deliveries in Asia. My task is to oversee the successful completion of the reorganization and subsequently further growth in the provided services.

Artificial intelligence is a frequently discussed topic, featured at every professional conference. However, many people still have only a superficial understanding of it. How can artificial intelligence help companies in logistics?

From a general perspective, artificial intelligence is a set of tools that learns from existing data and uses this “knowledge” to make decisions. In traditional software, the behavior of the program is written and coded explicitly, and no data is inputted. Artificial intelligence learns based on real data.

For the general public, the most visible part is Customer Support, where artificial intelligence is already often used. This expanded significantly with the advent of so-called generative artificial intelligence, or ChatGPT. That was a significant moment when artificial intelligence entered the awareness of the general public.

Currently, the most known applications are text understanding and generation. For instance, during order reception, where previously a human accepted orders by phone or email, artificial intelligence can now take these orders. It can accept an email with an order, read it, and enter it into the system, not only textually but also by voice.

If we look at logistics, the potential applications of artificial intelligence are truly numerous. It can, for example, predict the time when goods from a certain supplier will arrive. The client has a large number of suppliers who behave differently. Some send goods on time, others with a delay, and sometimes they don’t arrive at all. The client needs to plan stocking and destocking but also needs to know if the goods will be available at a specific time. They need a program that, based on historical data, predicts when which vehicle will arrive. In such cases, data is collected, fed into artificial intelligence, which processes it, and based on historical information, calculates what will happen in the future.

Another area where artificial intelligence can be applied in logistics is planning and scheduling. This can involve fleet management, planning which vehicle goes where, what it carries, which driver drives it, etc., so that the final solution is as efficient as possible. Finding the optimal plan is not trivial because even with a smaller number of vehicles, the combinations can be vast, and the human brain has limited capabilities compared to the combinatorial nature of artificial intelligence. Planning everything efficiently is one part of the task, and the other part is to plan it as quickly as possible in real-time. And artificial intelligence can handle such tasks very quickly. For this purpose, we have developed a tool called OptiSuite, which solves precisely these problems.

When it comes to planning, artificial intelligence can also solve shift scheduling in both logistics and activities related to logistics. Some logistics companies offer their clients additional services besides transport and warehousing. For example, for automotive manufacturers, they customize cars and need to allocate people in the workshop according to what each employee can do. Workforce Management is used for shift planning.

An area that I think is very important for logistics is the management of the entire supply chain, predicting and monitoring what happens within the supply chain, and when goods can be expected from which supplier.

There are indeed many ways to use artificial intelligence in logistics. For instance, in occupational safety, where artificial intelligence monitors the use of protective equipment and adherence to pedestrian zones.

However, I would also like to mention the part related to monitoring what happens in the physical world. Today, artificial intelligence can automatically count how many pallets are loaded into a truck or unloaded from it using a camera, as well as identify the type of goods. It can read license plates and recognize and record when which vehicle arrived and departed. This can be used in operations when planning and scheduling the receipt and dispatch of goods. It can suggest, for example, which gate the driver should bring the vehicle to or which queue they should join.

There must be countless interesting projects you have carried out. Do you have a specific example you would like to mention?

Here, I would like to mention a project we did for ŠKODA AUTO, where we optimized the process of loading containers. ŠKODA AUTO was looking for a way to optimize the loading process of transport containers to reduce unused space and thus the environmental impact of transportation.

The human brain was able to fill a container with a volume of 78 cubic meters with only 71 cubic meters. Finding new loading combinations for pallets became increasingly difficult, making it harder for the manufacturer to meet the loading criteria. ŠKODA AUTO, therefore, decided to look for a solution that would “expand” the possibilities of the human brain.

Together with Blindspot Solutions, we developed an intelligent system in the Microsoft Azure cloud ecosystem for optimizing the container loading process. In the project, we created an application that shows employees how to stack pallets. Now, using a tablet, they can find the best way to combine about 2,000 different pallet shapes into each 78 cubic meter container in just 30 seconds. The system takes into account the dimensions, weight, and material of the pallets and ensures proper weight distribution in each container. The application can also check if the correct pallets are loaded for transport every week.

This project saves ŠKODA AUTO approximately EUR 860,000 per year. But it’s not just about financial savings. Thanks to our solution, 300 fewer containers were shipped in the first year, representing five fully-loaded train sets and saving 160 tons of CO2 emissions.

It has been mentioned several times that quality data is the foundation. However, this can be a fundamental problem for many companies. What would you recommend to companies considering implementing artificial intelligence?

This brings us back to the connection between Blindspot Solutions and Adastra, because if a company wants to use artificial intelligence, it first needs to have quality data. In this case, it needs to build a data platform or data warehouse that allows it to collect the right data and store it in the right format so that artificial intelligence can work with it. This means that it is necessary to have what is called data governance, which is a way of managing data. And this is precisely what Adastra can help the customer with, what it has grown on, and in which it is a leader here.

Sometimes a client has a business problem and doesn’t know how to solve it. They may generally lack the data on which they could base qualified decisions. Among other things, Adastra can efficiently collect data in the physical world. We use smart sensors for this. Weconnect them to a smart platform and thus provide a tailor-made end-to-end IoT solution.

In general, it turns out that our clients don’t need to know in detail what technology will help them in business. They come to us with a specific problem, and after an initial analysis and assessment, we can devise the ideal solution for them and involve all our competencies. And as I said, clients in the entire global market can now access such complex services thanks to our partnership with Adastra. And that is our goal.

If a customer decides to implement artificial intelligence, how far back do they need to provide data for artificial intelligence to work with? And can artificial intelligence handle data that isn’t high quality?

The better the data quality, the better, of course. The basic period of data collection, when the data already has some value, is months. If a company has data that is in disarray but has been collecting it historically, we can clean up that data, organize it, and transform it in a certain way, i.e., clean it up. We then create a data repository, a data warehouse. And then we apply artificial intelligence to it. The data can then be used not only for artificial intelligence but also by management to work with it further. As they say, “What I don’t measure, I don’t control,” because if my data is not in order, I don’t see what’s happening in my company, and it’s then difficult to make decisions.

We also provide business consulting to companies, where we can connect to processes or recommend process adjustments precisely concerning data, its collection, and use. We can also recommend what additional data to collect. Today, it is common that some data can be artificially generated so that artificial intelligence can work with it. We know how to work with data from A to Z.

How are companies in the Czech Republic practically responding to the incorporation of artificial intelligence into their processes? When automation came along, there was a lot of talk about it, but in reality, companies were treading cautiously around it for quite a long time. I think that COVID then significantly accelerated its broader deployment in companies.

We see a similar scenario with artificial intelligence. We have been commercially involved in artificial intelligence since 2013, so we see how the readiness is gradually changing. At the beginning in 2013, companies were indeed not very open to AI. Today, companies’ attitudes are gradually changing as they begin to understand what artificial intelligence can bring them. We try to provide insight through our case studies. We provide companies with consultations, called AI Open Days. This means that we sit down with the customer and try to understand their business. Subsequently, we can generate use cases, including the recommended strategy, timeline, and expected benefits. That is the first, strategic part. The second part is technical and data related. Today, the application of artificial intelligence is entirely possible in the cloud. Thus, it may not restrict the company as much. If it already has a cloud solution, artificial intelligence can easily be added to it. If the company does not have a cloud, we create one. And the third part is how the company adopts the new technology.

We have encountered cases where the management approved the implementation of artificial intelligence, but there was some skepticism among employees. People who have been doing something for many years often cannot imagine that someone could do it better. Or they are afraid they will lose their jobs. But we can also work with that. We say that although artificial intelligence can function autonomously in some cases today, we are building a tool to help those people. It is not about artificial intelligence taking people’s jobs, but it should replace the routine steps that workers used to do, allowing them to specialize in activities with added value—for example, personal communication with the customer.

So, our activity is not just about the technical part, but also a lot about working with people. To explain the advantages and teach them to use these tools.

Artificial intelligence has been called a breakthrough technology with the potential to change the world. However, there are also certain concerns around it…

Artificial intelligence will undoubtedly have multifaceted applications in the future. People already use artificial intelligence in their daily lives, but they often don’t know it. For example, the internet or social media, where the content you see on social media is generated by artificial intelligence. When you log into the shopping app of your favorite retailer, very often, the offers you see are generated by artificial intelligence based on your previous shopping behavior.

Self-driving cars are a somewhat debatable topic. They are not yet fully autonomous, but many vehicles are already capable of driving for you to a large extent. You still have to keep your hands on the wheel, but on highways where lanes are marked, the most modern models already drive almost independently.

Artificial intelligence is also commonly encountered today in areas of text, image, and video generation. And it is precisely around these capabilities that concerns arise. Today, it is no longer difficult to generate a photo or video of myself where I speak to create a deep fake. Here we reach the societal implications and how artificial intelligence as a technology can be used for both positive and negative purposes. That worries me a little, but I believe we can face it. Probably again with the help of artificial intelligence.

Source: Logistic News