Bulgarian businesses are still not implementing AI effectively enough

Bulgarian businesses are using artificial intelligence (AI) but not to its full potential. At present AI is used primarily for optimization but not for enhancing processes. This became clear during a forum dedicated to the Bulgarian labor market and the impact of AI on it. During the discussion experts agreed that managers and entrepreneurs are using and implementing AI but there is still room for improvement.

Artificial intelligence is here in Bulgaria, but not quite – many companies still do not know what to use it for and have not integrated it into their processes,” said Tsvetelina Eremieva, Chief People Officer at Scalefocus.

According to her many companies are intensively testing where AI can be implemented, but even those conducting the tests lack the training to know exactly where to apply it. On the other hand there is also the issue of debunking the myth about AI: contrary to expectations not everyone actually likes or uses it.

Speed as a Challenge

Do not wait for everything to be perfect – start using the technology because it is changing incredibly fast,” added Dimitar Petrov, Managing Director of Phoenix Pharma IT Services.

He warned that the benefits of AI must be measured from the very beginning rather than starting with the technology itself. Petrov cited international studies showing that only 32% of companies can report a return on their investment in artificial intelligence.

Asparuh Vasilev, Senior Manager of Information and Cybersecurity at Nuvolo, highlighted a danger posed by speed, which is why security breaches now occur several times faster than before.

Five years ago it took an average of 21 d to exploit a vulnerability, whereas now it takes 7 – 8 h,” he noted as an example.

From a technological risk perspective nothing has changed compared to ten years ago, but speed is the key new threat, according to the expert. Vasilev highlighted the risk of so-called “shadow AI” and the danger of empowering artificial intelligence to make decisions on our behalf. His advice is for companies to work with a model that uses only their own data and to be careful about what information and access they grant it:

It is good to think of AI as a person we are bringing into the organization – it is important to be careful where we place it and what access badges we give it,” he summarized.

From Assistants to Quality Controllers

Another focus of the discussion was the changing role of humans with the advent of AI.

Alexander Papratilov, general manager of Wizer Technology, explained that until 3 – 4 years ago the technology was just beginning to take hold and functioned as an assistant and helper, whereas over the past year the roles have reversed so that AI does the work while humans merely monitor the quality of the work done.

We are becoming quality controllers – artificial intelligence does the work, and we oversee it,” he summarized.

Papatrilov also addressed the slow adoption of AI by Bulgarian businesses and the obstacles to this process. According to him there are two barriers to true implementation – a clear description of processes and access to high-quality data – as well as the psychological factor: despite predictions since 2022 that AI would replace humans, this has not happened.

Translated with DeepL.