Three out of five initial AI projects collapse at the start. Analyses by the MIT NANDA research network show that 95% of generative‑AI pilots did not produce measurable profit‑and‑loss results. In Poland, AI remains a key investment area for manufacturing firms, but only one in five is currently running pilot programs, the newspaper notes.
“Companies buy licenses, build pilots, train teams, and test agents. Users see higher productivity, management sees rising costs, and operating results often stay the same,” Jarosław Sokolnicki, co‑founder of exeAI, told Rzeczpospolita.
The main obstacles involve data quality, fragmented systems, and the absence of a clear strategy. An ABB report shows that the most common barriers include organizational and budget constraints (50%), regulatory requirements (41%), internal procedures (28%), and the lack of a long‑term plan (25%). Experts say firms often begin with technology instead of first improving processes and defining the business outcomes they want. People also remain a weak link. Although 72% of CEOs invest in new employee skills, companies still struggle to hire specialists in data integration, architecture, cybersecurity, and change management.
As a result, the era of announcing broad AI investments is ending. Boards now expect concrete metrics: higher productivity, lower costs, and faster workflows. A global Bain & Company study shows that 59% of CEOs believe their organizations achieve only a fraction of the expected results from AI transformation, Rzeczpospolita reports.
(pu)
Source: Rzeczpospolita