AI drives IT modernization – but legacy systems remain
A survey by MongoDB shows: Four out of five companies are modernizing their IT. However, the complete replacement of old systems often fails.
(Image: ZenitX/Shutterstock.com)
Four out of five companies in Germany are currently undertaking modernization projects for their IT infrastructure. This is shown by a survey conducted by database manufacturer MongoDB among 1,504 IT decision-makers. Artificial intelligence has emerged as the main driver: 43 percent of respondents stated that they are preparing their IT infrastructure for the use or scaling of AI.
In addition to AI integration, respondents also cited reducing operating expenses at 33 percent and more agile adaptation to operational and market-related requirements at 31 percent as important reasons. Almost three-quarters of the surveyed companies already use AI in some form; this figure rises to 80 percent for companies with more than 250 employees. AI is most frequently used in IT management (45 percent), customer service (42 percent), and data analysis and business intelligence (36 percent).
Companies primarily measure success by productivity increases (40 percent), customer satisfaction (38 percent), and error reduction (35 percent). Cost reduction ranks only fourth among success indicators at 33 percent. Only 4 percent of respondents indicated that they neither use AI nor plan to implement it.
Legacy systems continue to run in parallel
Despite the high number of ongoing modernization projects, the survey reveals a significant challenge: only one in three companies has been able to completely decommission its legacy systems. More than half of the respondents (52 percent) stated that cloud infrastructures do not match the performance of legacy systems. Other reasons for parallel operation included data migration problems (35 percent), the risk of operational disruptions (33 percent), and dependencies on critical processes (29 percent).
Many companies encounter structural limitations when expanding AI applications. Less than half of the respondents consider their existing data models to be a suitable foundation for AI applications; 14 percent consider them completely unsuitable. The biggest obstacles include a lack of data access (44 percent), insufficient real-time capabilities (44 percent), and unsuitable tools, data silos, and poor data quality (43 percent each).
Furthermore, almost half of the respondents indicated that a lack of relevant expertise hinders the implementation and scaling of AI technology. This shortage of skilled personnel affects both the technical implementation and the strategic planning of AI projects.
Videos by heise
Major differences by company size
The survey shows significant differences between small and medium-sized enterprises (SMEs) and large corporations. Companies with more than 250 employees are driving IT modernization and AI implementation much more strongly: 58 percent are already implementing corresponding projects, and 12 percent have completed them. This segment also invests more heavily in training (54 percent) and knowledge building.
At the same time, larger companies face greater challenges: 65 percent cited above-average integration complexity as a problem and are more frequently confronted with performance and security risks after migration. While SMEs show a similarly high interest in modernization, they have fewer financial and personnel resources, which slows down implementation.
Further information on the MongoDB survey can be found here.
(fo)
