Without Data and Business Challenge, AI Remains Only a Pilot Project

The role of a technology partner no longer ends with system delivery and a signature on the handover record – or at least it should not. Fiscalization 2.0, cloud migrations, ERP upgrades, and AI integrations – in Croatian companies, all these processes rarely take place separately, and even more rarely in the order that would be ideal from a technological perspective.

The discussion about Artificial Intelligence in the business sector is increasingly moving away from the tools themselves and toward their actual application and impact on operations. In an interview for ICTbusiness Media - ICTbusiness.info, Milan Listeš, General Manager of BE-terna Croatia, explains why AI should not be viewed as an isolated technological add-on, but as part of a broader redesign of processes and decision-making. The focus is on productivity, data quality, integration with existing systems, and the difference between pilot projects and solutions that deliver measurable results.

According to him, the business solutions market is entering a new phase in which it is no longer enough merely to digitalize existing processes, but rather to rethink them from the ground up. ERP, CRM, analytics, cloud, and Artificial Intelligence are increasingly merging into a single framework for business management, and users expect simplicity, speed, and measurable impact from these systems. At the same time, pressure is growing on companies to make better decisions based on data, with lower costs and greater operational precision. Such changes are also having a strong impact on the domestic ICT market, especially in the area of enterprise projects and digital transformation.

How do you currently see the role of Artificial Intelligence in the real reshaping of the way companies work?

Artificial Intelligence today has two very important roles: the first, still the primary one, is reshaping the work of the individual within the organization, and the second is, of course, reshaping the work of companies.

Given that LLMs have entered practical use through the mass market, the efficiency and productivity of every individual are the first to be affected by real transformation. Looking at the technologies we deal with, for several years now various generative AI functions have been an integral part of the Microsoft Office package, through which they significantly reshape individual productivity, but also of all the tools we implement on the market in the ERP, CE and Data segments, and in that way they are also reshaping the productivity of our consultants and developers.

On the other hand, reshaping the way companies work is a somewhat bigger challenge for every organization. It requires a certain level of organizational maturity, an understanding of business challenges, and only then the implementation of a specific technology. That is precisely why we put the definition of business goals first, and only then propose the technological realization of the project. A project structured that way results in business improvement, and not just the improvement of one isolated process.

How visible are measurable effects on productivity, cost reduction, and customer experience in your projects?

If the business challenge being addressed is properly defined, measurable effects are always visible. The segment in which we have the most experience is sales forecasting and the optimization of goods ordering and production planning. This type of project, in all industries that manage inventories, delivers a very exact and measurable return on investment, not only through productivity growth but also through a reduction in dead stock, positive cash flow, a lower number of stock-out situations, and so on.

How do you distinguish projects that deliver real business value from those that remain at the pilot-project level?

Many reasons influence whether a pilot project will come to life or not, but I would say there are two key ones. One is whether you have the data on which to anchor the technology, and the other is whether you have properly defined a real business challenge or are simply looking for a convenient place to test the technology. Technology without clear business value can hardly find its place in the industry.

How critical are data quality and integration with existing systems to the success of AI initiatives?

Data quality is crucial in every technological project that relies on data, whether it is an ERP project, a reporting project, or, especially, of course, Artificial Intelligence projects that use existing data to make some kind of prediction or generation.

On the other hand, I would say that integration with existing systems is particularly important in the early stages of a company’s transformation, when it is going through its first or second technology project, so that the process it is carrying out does not change too much. Our experience has shown that the level of AI adoption is much higher if it is well integrated into the existing process and does not significantly change employees’ daily activities, but simply enriches them with better information – in our example, this means that we integrate the proposal for automatic overnight goods ordering directly into the ERP from which the person would normally place the goods order. In that way, we do not significantly change the process, but simply give the user better information at their disposal and, through training, ensure the use of the solution.

Interest in AI and automation is growing strongly, but the focus is shifting to results. How do you see that on the market?

As with other transformations so far, there are those on the market who lead transformation and those who lag behind and wait for proven results. Likewise, today, some companies have more experience, perhaps even unsuccessful projects behind them from which they have learned something, and are now setting the prerequisites for new transformations in the right way. On the other hand, some companies may be in a dominant market position or in a specific industry that is not yet under so much pressure, and so they are still thinking about how and where to begin.

What business needs do clients most often want to address today with AI solutions?

Each industry has its specific features, but I will try to cover most of them in several categories. The examples I have mentioned so far fall into the category of increasing efficiency and reducing costs: planning optimization, predictive maintenance in manufacturing and energy, or, for example, waste reduction in the fresh food segment, especially in retail. The second category in which clients seek the application of AI is personalization and revenue growth: in sectors such as banking, telecoms, or retail, work is being done on examples of personalized offers, product recommendations, or sometimes price optimization. And the third I would mention for now, although there is a much broader range of applications on the market, is knowledge and integration automation: with the emergence of generative AI, strong interest has appeared in internal AI assistants – document and procedure search, for example, in service organizations, or the automation of responses to clients in industries with intensive customer service.

How is the role of technology partners changing today compared with traditional IT suppliers?

The role is certainly changing, but like any transition, this one is not simple for IT suppliers either. For the market to see an IT company as a technology partner, several components are required, and the minimum is a complete understanding of the industry. Not only process knowledge of, for example, the “order to cash” process, but real knowledge of the industry, what the key KPIs are, what the main margin generators are in that industry, what the regulatory or technological constraints are, and so on. The second point I would highlight, if a company possesses this breadth, is participation in defining the business case that is being modernized, because only by going deeply into the issue and understanding how the user uses technology daily can quality implementation be ensured.

This is particularly visible when introducing solutions based on AI, because success requires an understanding of data quality and its sources (ERP, CE, production lines, etc.), the importance of data architecture, modeling, understanding the industry, interpretation of results, education, and work with users, and so on.

How do you build trust and long-term partnerships in complex projects? Which competencies today most distinguish a quality partner from the competition?

In my opinion, several elements form the foundations of long-term partnerships, which in the end also make the difference in the market. Along with the domain knowledge we mentioned, the ability to understand the technological process end-to-end is also important, but not in the sense that as one organization we can offer everything, but rather the opposite – that you have a network of partners and that you can recognize that not everything needs to be the subject of custom development or customization, but that somewhere there are specialized companies on the market that may be a better fit for your client. In my experience, this type of knowledge, translated into transparent communication with the client, forms one of the foundations of a long-term relationship.

The second segment is industry knowledge, on which we work continuously, not only through the development of our consultants through project experience, but also by bringing in experts from the key industries that are in our focus. It is precisely the combination and synergy of these different perspectives and experiences that create key differentiation in the market.

And the third segment is that for us, the long-term relationship does not end after implementation. We continue to support the client after go-live and ensure that the solution delivers the planned business value.

BE-terna has had a long presence on the market. How has your core business changed over the years?

In business development, every innovation that you try to introduce into operations can be approached through build, buy, or borrow, which means that you can build the expertise through internal resources, buy it through the acquisition of a company, or borrow it through a partnership. BE-terna has grown through a combination of all three approaches. Where we needed to own the knowledge, we built it internally – such as industry knowledge, implementation excellence, and so on – and where it was logical, we relied on cooperation with global technology partners and in that segment “borrowed” best practices. Looking back, we have developed from an ERP implementer into a strategic partner for digital transformation, and it is precisely that partnership that today forms the core of our business. Companies choose us not only because of our technological expertise, but because of our ability to understand their business challenges and support them over the long term through processes of change and growth.

How did you move from ERP implementations to complex digital and AI solutions?

This is a segment that we gradually built internally. At a certain point, the development of a reporting system was logically added to the ERP business, but not in a narrow form as an extension of the ERP solution, but as a comprehensive solution that today is based on understanding the importance of data quality, data management, warehouse construction, and business analytics. This step proved to be a key prerequisite for the development of complex digital solutions based on Artificial Intelligence, because it is very difficult to offer any kind of advisory work in the Artificial Intelligence segment without understanding the flow of a single piece of data from source to use.

What was crucial for the development of the team was a senior team lead who built a team around himself and, of course, the implementation of the first project, during which we went through quite a few teething problems, but at the same time also understood the full potential of using AI specifically in the segment of sales forecasting and inventory management.

How does experience from different sectors affect the development of new services?

Two trends are very quickly changing the role of technology partners. Over the past ten or so years, technology has largely become a commodity from the perspective of the end user. A person who, in private life, uses Wolt, Uber, or Netflix as an end product, or who goes through the “sales process” of Revolut or some other digital banking application, comes to work with such expectations and more or less expects a similar experience in the world of business software.

At the same time, that also means that the gap between the technological knowledge of the client and the partner is narrowing, as technology becomes ever more accessible. These two trends have greatly accelerated the process by which BE-terna, like other technology partners, must acquire industry expertise. We manage this process through two segments: one is defining repeatable implementation teams by industry, to the extent possible, and the other is bringing in consultants and developers who possess industry knowledge and have worked in accounting, sales, or production controlling, in the introduction of business solutions on the client side, and so on.

How does membership in the international Telefónica Tech group affect your business in Croatia?

Of course, being part of a global organization opens business opportunities, for example, in segments in which our region has a center of expertise within the BE-terna group. Still, from the acquisition until today, our focus has been and remains full support for business operations for companies in Croatia.

How do you view Fiscalization 2.0 and other regulatory initiatives from the implementer’s perspective? What are the biggest technical and organizational challenges of adapting systems?

All regulatory initiatives, by nature, create somewhat greater pressure both on clients and on us as implementers. Challenges in such extensive changes are always expected, and several months are needed for the system to fully settle in – we see that as feedback from the rest of the market as well, not only from our clients. While the implementation process itself is ongoing, I think clients do not think much about it as an opportunity for modernization, but once impressions settle, it becomes clear that companies on newer, more modern system versions have gone through this process more easily.

How do you see the development of AI and business platforms in the medium term? Which technologies will have the greatest impact on business? Will AI become a standard part of ERP systems? How should companies prepare for the upcoming changes?

In the context of Artificial Intelligence and its impact on business solutions, an extremely exciting period lies ahead. At the moment, business applications are arranged in several separate layers that communicate through different integrations, but the fact is that large companies have a very large number of disconnected systems.

I believe that the “pieces” of Artificial Intelligence that we see today in every system are only an announcement of the real transformation that awaits us. In that transformation, the foundations will continue to be transactional systems (ERP, CRM, or various industrial applications in manufacturing), but the demands placed on the cleanliness and quality of data will be extremely high, and of course, those systems will have a large number of functionalities based on Artificial Intelligence, primarily in the segment of productivity growth for users.

The second layer represents a “Governed Data Backbone” – something that we have until now viewed as a data warehouse or reporting system will now become a centralized place for collecting and cleaning data, with a strong emphasis on data cleanliness and quality before it is consumed somewhere.

And the final key layer represents something that we can imagine as an “AI thinking engine,” that is, a layer that will be able to make decisions independently (or with the help of users, depending on the application) and revise its behavior based on what it has learned. The user interface for such systems could be different environments, from a reporting tool to chat platforms (Teams, WhatsApp), and so on. It is important to note that such concepts are not entirely new; already today in the region, we have companies that order more than 80 percent of their retail assortment automatically without a human even looking at it. Still, today we talk about that as isolated examples, while I believe that in the coming period, technology will enable that way of doing business and making decisions to become a standard in most organizations.

How do you see the role of BE-terna in that transition?

If you look at these three components, it is clear that these are segments of the business that BE-terna has been dealing with for many years. Our deep understanding of transactional systems, data technologies, and experience in successful AI implementations puts us in a leading position to help companies connect isolated silos into a single, modern management system that delivers measurable business value.