Only a handful of generative AI projects make it into production – here’s why
Many generative AI projects are failing as firms struggle with issues around data quality, expertise, and technology, according to a study from Informatica.
Less than two-fifths (38%) of AI projects have successfully transitioned into production, the report revealed, while 67% of firms have seen less than half of their generative AI pilots reach production.
Similarly, over two-fifths (41%) slowed, paused, or entirely halted generative AI initiatives over the past year, though this was reportedly the minority – 46% maintained a steady pace, and 13% moved at speed.
The study fielded responses from 600 data leaders, including chief data officers (CDOs) and chief data analytics officers (CDAOs).
Businesses hitting a wall largely blamed issues of data and technology. Nearly half (43%) cited a lack of data quality and readiness as the biggest obstacle to reaching production, while 43% also blamed a lack of technical maturity. In the UK and the EU, 48% blamed technology compared to 37% in the US.
Data reliability is a major barrier for firms that have completed some of their generative AI pilots, with 56% describing it as one of a few key barriers.
Similarly, 10% indicated it’s a top barrier to moving into production.
Many of these challenges have remained persistent since 2023, Informatica said. Quality of data and data privacy and protection are still at the top of the list when it comes to data-related challenges for businesses, as they were the year before.
Organizations slowing down or halting their AI projects may do so because of increased difficulty in justifying the investment. Just under half (46%) say that cybersecurity and privacy issues have made demonstrating the value of generative AI difficult.
There can also be a lack of trust, with 43% expressing concern over the reliability of generative AI results.
“The pressure to demonstrate ROI, combined with concerns over data quality and regulatory delays, is creating a perfect storm for data leaders. Unless their data is ready for AI, they won’t be able to unlock generative AI’s full potential,” Emilio Valdes, SVP at Informatica, said.
“To succeed, organisations must focus on strengthening their data foundations. Ensuring data is accurate, holistic, integrated, up-to-date and well-governed is a critical step in converting pilots to transformational projects, proving value and navigating compliance challenges,” Valdes added.
Source link
