Ubiquitous Yet Uncontrolled AI: 96% of IT Professionals Affected

⚡

Key Takeaways

1A study reveals that 96% of IT professionals use AI, but only 49% do so frequently.

251% of professionals believe that AI reduces their productivity, requiring 10 hours weekly for data validation.

3The main obstacles to AI include interpreting results and a lack of analytical skills.

💡Why it matters — The widespread use of AI highlights a crucial need for skills in data validation and interpretation to optimize its effectiveness.

The Massive Use of AI by IT Professionals

A global study conducted by Alteryx, involving 700 data analysts and 700 IT leaders, highlights the extent of artificial intelligence usage in the tech sector. Indeed, 96% of the professionals surveyed report integrating AI into their activities. However, only 49% of them are regular users, utilizing AI all or most of the time.

In the next 12 months, 59% of respondents plan to actively use AI agents. Additionally, 50% are willing to grant unlimited access to their data to these agents, although 44% emphasize the importance of human oversight. The security implications related to this unlimited access were not discussed in the report.

Common Applications of AI

The study identifies several common applications of AI among IT professionals. Writing standardized communications or summaries for stakeholders is cited by 59% of participants. Planning or routing workflow tasks, such as alert triage and process automation, is mentioned by 54%. The generation of reports or standard dashboards without manual intervention is utilized by 48% of professionals.

Other applications include monitoring key performance indicators and triggering alerts or actions (45%), as well as cleaning, preprocessing, or validating routine datasets (45%). Performing routine statistical analyses or basic predictive modeling is practiced by 34%, while the automatic generation of insights or recommendations from data concerns 23% of respondents.

Time Spent on Data Work

"Fundamental data work," which includes cleaning and preparing data for ingestion by AI models or augmented generation platforms, occupies a significant portion of data analysts' time. Respondents report spending nearly six hours per week on these tasks, with 48% dedicating between six and ten hours weekly.

51% of professionals believe that working with AI reduces their productivity, requiring about 10 hours per week for data preparation and validation.

Tools and Barriers to AI Implementation

To manage this data work, professionals primarily use spreadsheets (61%), business intelligence tools (56%), and dedicated data preparation platforms (51%).

The barriers to AI implementation are numerous. The difficulty in interpreting or explaining AI results to decision-makers is mentioned by 55% of respondents. The lack of analytical skills among business users is a problem for 54%. Furthermore, 50% of participants report that the data is not sufficiently clean, integrated, or governed.

Other barriers include a lack of clarity regarding ownership or accountability for decisions (49%) and technical limitations of AI tools or infrastructure (45%).

Validation of AI Results

Validating the results generated by AI is also a time-consuming task. Analysts in the survey spend nearly four hours per week validating or correcting these results. One in six respondents even dedicates an entire workday, six hours or more, to adjusting AI results. When adding the six hours spent on fundamental data work, this represents an "AI tax" of nearly two days per week on professionals' time.

This situation highlights an emerging skill that is becoming increasingly valuable in the age of AI: the validation of AI results. Although AI can accelerate work, organizations still need human oversight to ensure that the results are consistent, explainable, and trustworthy.