In fact, 72 per cent of enterprises admitted to facing significant challenges with data quality and scaling data practices, having a significant impact on the output quality of AI models.
Out of the 750 IT decision makers surveyed, a quarter (24 per cent) said that they had implemented generative AI at scale, with employee productivity tools and customer services tools the most common implementation.
“Our report highlights a concerning trend: many enterpirses, in their eagerness to harness AI, overlook the need for a solid foundation. This oversight not only diminishes the effectiveness of their AI solutions but also exposes them to a multitude of security threats,” said Kunal Anand, CTO at F5.
The report also highlighted key barriers to scaling AI technology within the data layer, with 72% of respondents pointing to data quality and the inability to expand data practices as major issues.
Additionally, 53% of respondents identify a lack of AI and data-related skillsets as significant obstacles.
When it comes to the infrastructure layer, enterprises express concerns about the cost of computing resources (62%), model security (57%), and overall model performance (55%).
Commenting on the findings, Roman Kucera, CTO and Head of AI, Ataccama: “Before deploying AI, enterprises must ensure the quality of their data otherwise poor data will impact the accuracy of output. Insufficient or poor quality data with errors and anomalies will undermine the trustworthiness of AI-based insights. We see companies that have prioritised data quality in the past becoming the ones making the leap to AI most easily today and enjoying fast time to value. For those starting out on the journey, investing in better AI training, realising operational efficiencies and unlocking data for business users will all facilitate quicker ROI for the business as a whole."
Another recent survey from Lanop Business and Tax Advisors highlighted that 66% UK of IT decision makers feel their businesses is lagging behind in AI development. At the same time, 92% of enterprises claim to be prioritising data quality as essential when feeding data into AI models.
