Key Takeaways:
– Data quality is identified as the main barrier to implementing generative AI (GenAI), based on a survey of data leaders by Informatica.
– 45% of companies have already adopted GenAI, and 53% plan to do so.
– Data privacy and protection, AI ethics, the volume of training data, and AI governance are other GenAI challenges.
– All survey participants are investing in data management capabilities.
– A large number of companies are utilizing more than 1,000 separate data sources.
– Informatica’s survey participants are operating five or more tools for their data management work.
The Success of GenAI Tied to Data Quality
Informatica’s latest CDO Insights 2024 report details the perspectives of 600 data leaders at large global companies on GenAI. It reveals that 45% of these companies have implemented GenAI in some capacity. A further 53% plan to do so, with 36% intending to institute it within two years. It’s a striking outcome for a technology that surfaced only about 14 months ago and illustrates just how rapid the adoption of GenAI has been.
Data quality, however, emerges as the most significant barrier to successfully installing GenAI. Despite current artificial intelligence (AI) and machine learning (ML) technologies being simpler to deal with than those of the past, excellent data is a vital component. Poor data quality not only completes the failure of a GenAI project, but also undermines all types of AI or ML projects.
Data Leaders’ Top Concerns for GenAI Success
The survey highlights the concerns of those deploying or planning to deploy GenAI. A total of 42% of data leaders, or approximately 588 out of the 600 respondents, view data quality as the number one concern. This is followed by data privacy and protection, AI ethics, the quantity of data for training and fine-tuning language models, and AI governance.
Spurring GenAI Success Through Advanced Data Management
Given these concerns, investment in data management capabilities is on the rise. Interestingly, all survey participants signal a commitment to investing in these capabilities to support their data strategies and priorities. This decision could be a game-changer in addressing data quality issues and promoting GenAI adoption.
Yet, managing data is not a monolithic task. According to Informatica’s findings, 58% of those surveyed are using at least five different tools for their data management operations. This is despite nearly half of the respondents reporting that these tools were not available as cloud-hosted services.
The Challenge of Managing Multiple Data Sources
A large factor plaguing GenAI initiatives is the sheer number of data sources companies must handle. Two out of five firms admit to dealing with 1,000 or more data sources. The projection for 2024 is not much better, with almost 80% of survey respondents expecting their number of data sources to increase.
Navigating Forward: Priorities for 2024
This year, a key priority for 39% of data leaders is to amplify the reliability and consistency of data for GenAI use cases. Similarly, fostering a data-driven culture and improved data literacy is also deemed a top priority by the same percentage of leaders.
The latter emerging as a priority isn’t surprising to Jitesh Ghai, Informatica’s chief product officer. He asserts that the implementation of GenAI and the necessary data strategies continue to dominate most data leaders’ bandwidth. Therefore, investments in highly integrated data management capabilities seem to be the deciding factor in how successfully companies can unlock GenAI’s full potential.
As we look to 2024 and beyond, it becomes increasingly clear that organizations need to tackle data quality and data management issues head-on. Both are critical pillars for GenAI’s successful integration and their neglect might stall AI’s broader adoption in the corporate landscape.
