Poor data quality is hindering AI adoption

A new report reveals that although 84 percent of IT leaders say a Configuration Management Database (CMDB) is essential to driving decision-making and operations, a majority feel their current systems lack the data quality, accuracy and completeness they need, hindering the ability to maximize enterprise AI implementations.

The study from Device42 polled IT leaders across industries including finance, healthcare, government, and technology, and finds that over 50 percent of respondents use CMDBs, monitoring tools, or manual discovery processes, to gain insights about their infrastructure, yet 58 percent report a lack of confidence in their visibility.

“The message is clear. IT teams understand the value of having comprehensive visibility, but the challenge is that many still rely on fragmented, outdated, or manual systems that can’t keep pace with today’s complex and fast changing environments,” says Raj Jalan, senior vice president and general manager of the Device42 business unit, of which he is also the founder. “What they need is a solution that delivers automation, accuracy, and real-time insight across hybrid infrastructure.”

Among other findings, just 17 percent of respondents say their CMDB is fully accurate and used regularly. Gaining visibility into complex environments and reducing operational costs is ranked among the top IT challenges.

All of which is holding back AI adoption, 64 percent of IT teams say they have not adopted AI, with data quality and cybersecurity risk cited as the two biggest concerns. 45 percent of respondents expressed interest in adopting AI, but only if data quality and risk thresholds can be addressed.

As organizations seek to integrate AI into their infrastructure and operations, the effectiveness of those systems depends entirely on the quality, completeness, and trustworthiness of the underlying data. Without accurate insight into asset relationships, usage patterns, and system configurations, AI tools are left to operate with blind spots, introducing risks instead of efficiencies. “You can’t modernize what you can’t see,” adds Jalan. “And you certainly can’t trust AI without trust in your data.”

You cab read more on the Device42 blog.

Image credit: Khakimullin/depositphotos.com