96% of IT pros use AI now: Their top 7 agentic applications and biggest implementation roadblocks
Original reporting by ZDNet
The world of data analysis is abuzz with artificial intelligence, yet a new global survey of 1,400 data analysts and IT leaders reveals a nuanced reality beneath the hype. While a staggering 96% report using AI in their work, less than half can be considered frequent users, challenging perceptions of widespread deep integration. This initial enthusiasm, however, is met with practical challenges and surprising steadfastness in traditional tools, with spreadsheets still dominating essential data preparation.
The Agentic Shift
Looking ahead, a significant transformation is on the horizon with agentic AI. Nearly six in ten professionals anticipate actively deploying AI agents within the next year, with fully half expressing a willingness to grant these autonomous systems "unrestricted access" to their data. This readiness, while promising profound efficiency gains, simultaneously underscores the critical need for human oversight—a point highlighted by 44% of respondents, especially as security implications remain largely unexplored.
Despite AI's growing footprint, "foundational data work" remains a persistent time sink, with professionals spending an average of six hours weekly cleaning and preparing data. This, coupled with another four hours spent validating AI-generated outputs, amounts to nearly two full workdays dedicated to AI-related data tasks each week. This significant time investment, an "AI tax," reveals that while AI can accelerate workflows, human input is crucial for ensuring trusted, explainable outcomes, and interpreting complex AI outputs remains a major barrier to deeper business integration and real-time decision-making.
The recent Alteryx survey offers a nuanced portrait of AI integration within IT and data professions. While adoption is nearly universal, its depth remains inconsistent, with many professionals still spending significant time on foundational data preparation and output validation. This "AI tax"—almost two days per week—reveals that current AI implementations are augmenting rather than fully automating existing workflows, often layering onto traditional tools like spreadsheets. The imminent surge in agentic AI, coupled with a willingness to grant it unrestricted data access, underscores a pivotal moment. While these agents promise unprecedented efficiency in tasks from scheduling to reporting, the emphasis on human oversight signals a critical awareness of potential risks and the ongoing need for accountability.
Evolving Professional Landscape
Looking ahead, the findings illuminate a future where data professionals' roles evolve from execution to curation and validation. The persistent challenges in interpreting AI outputs and the scarcity of analytical skills across businesses highlight the growing importance of human expertise in sense-making. Building trust in AI outputs, ensuring their explainability, and fostering sophisticated analytical capabilities will be paramount for organizations seeking to leverage AI's full potential. The journey isn't merely about deploying AI tools; it's about strategically integrating them with human intelligence to navigate complexity, ensure accuracy, and drive meaningful, trusted insights in an increasingly AI-driven world.