Azernews.Az

Qabil Ashirov

Read more

The primary reason artificial intelligence (AI) projects fail to move into production is related to data challenges, according to a senior official of the Central Bank of Azerbaijan, AzerNEWS reports.

Speaking at the “X International Finance and Banking Summit,” Mahir Guluzade, Head of the Data Engineering Division at the Data Department of the Central Bank of Azerbaijan (CBA), said that AI technologies are evolving rapidly and the era of artificial intelligence has already begun.

“Since 2025, we have been witnessing more significant advancements in the field of artificial intelligence. Agentic AI has already become part of our daily lives. It is expected that 2026–2027 will mark the fastest phase of development in the AI era,” Guluzade noted.

However, he emphasized that the financial sector continues to face major challenges in implementing AI solutions, particularly due to the lack of high-quality, production-ready data.

“According to Gartner research, many AI products and enterprise solutions fail to reach production environments. One of the key reasons is the lack of ‘AI-ready’ data,” he said.

Guluzade added that many organizations assume that once they build a data warehouse or a data lakehouse, they are ready to deploy AI projects. However, he stressed that data collection alone is not sufficient.

“The data must be meaningful for AI systems; they should be able to correctly understand and interpret it. Therefore, a ‘data-ready’ approach designed only for analytics is not enough. For AI solutions to function effectively, it is essential to create a semantic context layer that adds meaning to the data,” he explained.