According to the study How Polish Companies Implement AI, conducted by CubeResearch on behalf of EY among 499 medium-sized and large Polish companies in the last quarter of 2025, as many as 49% of Polish enterprises were disappointed with their artificial intelligence implementation over the past year.

“The figures from the EY study sound dramatic only from one perspective: 49% of companies are disappointed and 17% would not repeat the implementation. On the other hand, the same report shows that 51% of companies achieved or exceeded their AI-related goals, while 53% reduced costs thanks to artificial intelligence,” notes Krystian Sperka, VP & Managing Director at Tenarai Europe. He adds: “The problem therefore does not lie in the technology itself, but in the sequence of actions taken.”

The most common mistake observed in many companies—not only in Poland—is purchasing an AI model before anyone has asked which business process it is supposed to improve.

“I agree with the opinion expressed by EY experts that the starting point must be a specific operational problem: response time, unit cost, the number of errors on a production line, or the decision-making path in the back office. In the data & AI field, you do not sell a model or technology. You sell a tangible business outcome. Only then does AI cease to be, as EY Consulting’s Iwona Kozera described it in one of her media statements, an ‘expensive gadget,’” Sperka stated.

The second reason for disappointment is neglecting the foundation. Fewer than 40% of Polish companies have a well-organized data infrastructure that enables them to move beyond off-the-shelf tools. This is the real diagnosis: without integrating source systems such as ERP, MES, and telemetry data, and without an analytical layer, even the best model remains a laboratory experiment.

“In our projects, 60–70% of the first months are spent working with data, not with the model itself. This may not make for an attractive boardroom slide, but it is precisely where the difference lies between a pilot project and a production deployment,” emphasizes Krystian Sperka.

Finally, change does not have to mean revolution.

“For companies that have become stuck at the third level of AI maturity, I usually recommend the opposite approach to the hype: add an agent or model to an existing, functioning application and measure the impact within a single, narrowly defined process. Today, the greatest return on AI comes not from software development itself, but from business operations and areas where domain expertise meets technology. We are no longer a traditional IT provider. We are becoming a business transformation partner built around AI agents—solutions that not only analyze data but also independently take actions within our clients’ processes,” concludes Sperka.

Data as the Foundation. AI and Agents as the Lever

For years, Tenarai’s Kraków team has been developing expertise in data & analytics and AI—and this is where the most significant shift in client engagement is currently taking place. Projects typically begin with organizing data sources: integrating production systems, ERP, MES, and telemetry data from technological production lines. Next comes the creation of an analytical layer on which predictive models and AI solutions can be built. Without this foundation, company experts emphasize, even the best model remains a laboratory experiment.

“In the data & AI domain, we deliver measurable business outcomes to our clients: faster decision-making, fewer errors, lower operating costs, and better utilization of people where their expertise is truly needed. AI agents are currently the most natural way to deliver these outcomes directly within the client’s processes rather than alongside them,” explain Tenarai experts.

An important part of this offering is Ignis, the company’s proprietary platform for designing, testing, and deploying AI solutions, including entire teams of collaborating AI agents. The platform reduces the time from idea and initial proof of concept (PoC) to production deployment from weeks and months to just days.

Tenarai Europe (formerly Infogain Poland) has operated in Poland since 2020, when it launched its Kraków technology center, which also became the company’s first hub in Poland and Europe. Since then, it has served as a nearshore center for clients across Europe and the United States, while also becoming a key center of excellence in AI and software engineering. The Polish company actively participates in projects in countries including Romania and Uruguay.