Data quality a major barrier to AI success

With the rapid adoption of AI across industries, nearly one in two (42%) UK companies identified data as their top concern when implementing AI projects, but few IT leaders are taking steps to ensure proper data quality and management, jeopardising the success of AI initiatives according to a new survey from Hitachi Vantara. The Hitachi Vantara State of Data Infrastructure Survey reinforced the critical role that data infrastructure and data management can play in terms of overall data quality and the ability to drive positive AI outcomes.

“Using high-quality data” was the most common reason provided for why AI projects were successful both in the UK and globally, with 41% of UK respondents in agreement. However, AI has led to a dramatic increase in the amount of data storage that businesses require, with the amount of data expected to increase 150% by 2026. As a result, storing, managing and tagging data to ensure quality for use in AI models is getting harder.

The company commissioned the global survey of 1,200 C-level executives and IT decision-makers across 15 countries. The survey found that most businesses were focused on security risks at the expense of data quality, sustainability and infrastructure management. Key UK findings include:

Despite the importance of data quality in AI success, 56% of UK IT leaders said more than half their data is dark.

Few are taking steps to improve their data quality; only 40% are enhancing training data quality to explain AI outputs and 20% don’t review datasets for quality.

45% of businesses are also experiencing data storage concerns with AI on the rise. The average large organisation globally now manages approximately 150 petabytes (PB) of data, a figure projected to exceed 300 PB by the end of 2026—enough to store every film ever made since 1950 nearly 200 times over in 4K resolution.

AI strategy is lacking ROI analysis or sustainability considerations, as only 44% ranked sustainability as a priority in AI implementation. A similar number (42%) said they were prioritising ROI.

Security is top-of-mind due to the risks presented; 70% acknowledge that a significant data loss could be catastrophic to their operations, while 65% of respondents are concerned that AI will provide hackers with enhanced tools.

“The success of AI hinges on trust—trust in the system and in the output,” said Sasan Moaveni, Global Business Lead for AI & High-Performance Data Platforms at Hitachi Vantara. “If those early interactions with AI don’t go well, it leaves a lasting impact on how organisations approach it in the future. A lot of companies are diving into AI without a solid strategy or proper training simply to keep up, but this can backfire. Successful AI projects start with a clear plan—defined use cases, desired outcomes, and infrastructure built to handle massive data responsibly. Resiliency and energy efficiency are key. Without sustainable infrastructure, businesses may find themselves facing costly overhauls to meet future regulations. Getting AI right isn’t just about innovation; it’s about building a trustworthy and future-ready foundation.”

Why Data Infrastructure is Key in Driving AI Success

Despite recognising data quality as the top concern for successful AI (42%) many organisations lack the infrastructure to support consistent data quality standards. More than half (56%) are testing and iterating on AI in real-time without controlled environments, leaving room for significant risk and potential vulnerabilities. Only 12% report using sandboxes to contain AI experimentation, which raises concerns around the potential for security breaches and flawed data outputs. Modern infrastructure offers a solution, as it is designed to be more energy efficient, allowing organisations to improve performance while also reducing their carbon footprint. By adopting sustainable, cutting-edge infrastructure, businesses can enhance data quality, mitigate risks, and support environmentally responsible AI growth.

“Companies want to work with partners that help them grow, help them be more efficient or reduce and mitigate risk. We are addressing risk,” said Octavian Tanase, Chief Product Officer, Hitachi Vantara. “We are providing automation which translates into operational simplicity, so companies are more efficient. If companies get more insights out of the data, that will help them compete and grow. The failure to deploy robust infrastructure for data quality and testing undercuts AI’s potential, making it essential for organisations to prioritise a solid data foundation before scaling AI initiatives.”

Having a Trusted Partner Can Help

Additionally, the survey reveals that as organisations advance AI initiatives, most IT leaders recognise the need for third-party support in critical areas, including:

Hardware - To be effective, hardware needs to be secure, available 24/7, and efficient to meet sustainability goals. In the survey, 27% of IT leaders report needing assistance to create scalable, future-proof hardware solutions.

Data Storage and Processing Solutions - Effective data solutions bring data closer to users while emphasising security and sustainability. The survey found that 26% of leaders need help with ROT data storage and data preparation, while 37% seek assistance with data processing.

Software - Secure, resilient software is vital for protecting against cyber risks and ensuring data accessibility. 35% of IT leaders require third-party expertise for developing effective AI models.

Skilled Staff - The skills gap remains a hurdle, with 46% of leaders building AI skills through experimentation and 42% relying on self-teaching.