Follow ZDNET: Add us as your preferred source Google.
Important points of ZDNET
- Almost all data and IT professionals use AI, but few are heavy users.
- Many will grant AI agents unrestricted data access.
- Preparing and validating AI data takes approximately 10 hours per week.
If you want to know what’s going on in the artificial intelligence storm, see what data analysts around the world are up to. Of course they’re bullish on AI, but they’re still using spreadsheets and only a handful are working with real-time data.
That’s the word from Alteryx’s new global survey of 700 data analysts and 700 IT leaders. Although 96% report using AI in their work, only half consider themselves to use AI tools frequently, and 49% report using AI all or most of the time.
Also: 51% of professionals say AI is hurting productivity – two steps to stop it
Agent AI is high on the agenda, with nearly six in 10 (59%) respondents predicting they will actively adopt AI agents within the next 12 months. Furthermore, at least half said they would be willing to give AI agents “unrestricted access” to their data.
Although the security implications of such access were not discussed in the survey report, 44% said it was important to include human oversight as part of such access.
Most common agent AI applications
The most common agent AI applications in operation today are communication drafting and workflow scheduling.
Where AI agents come into play:
- Drafting standardized communications or summaries for stakeholders: 59%
- Scheduling or routing workflow tasks such as alert triage and process automation: 54%
- Generate standard reports or dashboards without manual intervention: 48%
- Monitoring key performance indicators and triggering alerts or actions: 45%
- Routine data set cleaning, preprocessing, or validation: 45%
- Routine statistical analysis or running basic predictive models: 34%
- Automatically generate insights and recommendations from data: 23%
“Fundamental data work”—cleaning and preparing data for ingestion by AI models or related search enhancement generation platforms—still takes a significant amount of data analyst time. Respondents reported spending nearly 6 hours per week on such tasks, with 48% spending between 6 and 10 hours per week. The tools they use to handle these tasks are spreadsheets (61% said), followed by business intelligence tools (56%) and dedicated data preparation platforms (51%).
Also: Build an agent AI strategy that delivers results without risking business failure
“The continued dominance of spreadsheets reflects a broader reality,” the study’s authors suggest. “AI is layered on top of existing workflows rather than replacing them.”
Another surprising finding is that despite all the attention paid to real-time responsiveness, very few organizations actually have real-time capabilities. Only 20% report that they can move from data analysis to business decisions within hours, and only 5% say they support real-time decision making.
What are the biggest barriers to AI?
Respondents said they explain AI outcomes to business decision makers. There is also a significant lack of analytical skills across the company.
Barriers to AI in business decision making:
- Difficulty interpreting or explaining AI output to decision makers: 55%
- Business users have limited analytical skills: 54%
- Data is not sufficiently clean, integrated, or managed: 50%
- Lack of clear ownership and accountability for decisions: 49%
- Technical limitations of AI tools or infrastructure: 45%
Generating insights from AI is never a one-time task, and it can also waste data analysts’ time. Analysts in the study spend nearly four hours a week validating or modifying AI-generated output. One in six people say they spend nearly their entire workday (six hours or more) tinkering with AI results. Add in the six hours spent on basic data work cited above, and the AI “load” adds almost two days a week to the expert’s time.
Also: More than 80% of US government agencies are already using AI agents – and that’s just the beginning
This points to a new skill set that is becoming increasingly valuable in the AI era: validating the output of AI. According to the study’s authors, this “suggests that while AI can accelerate operations, organizations still need human oversight to ensure results are consistent, explainable, and reliable.”
