What GA’s new survey reveals about skills gaps, productivity gains, and the future of AI in data work
A shocker to no one: AI is rapidly reshaping the data analytics landscape.
But more intriguing: According to our new survey research, it’s transforming workflows faster than organizations are transforming skills. That mismatch is pushing data professionals toward risky behaviors, widening confidence gaps and creating uneven productivity outcomes across the field.
We recently surveyed 269 data professionals across the U.S. and U.K., including analysts, data scientists, data engineers, and database administrators.
The spoiler? While adoption of AI with data pros is high, the results reveal a critical truth: AI succeeds only when teams are trained to use it in ways that are secure, strategic, and aligned to their roles.
Below, we break down some of the other important insights that came out of our survey. Let’s get into it.
AI is everywhere—and many data professionals are going rogue to use it
AI adoption among data teams is nearly universal:
- 79% of data professionals use AI at work
- 52% use AI agents
- The typical data professional uses AI 6 times per day
But beneath the surface, there’s a significant compliance red flag:
- 40% admit to using unapproved AI tools
- 17% say they primarily rely on free, publicly available tools
- Many of these professionals handle sensitive or proprietary data
It should be deeply concerning to business leaders… This is what happens when technology adoption outpaces training and skills.
— Daniele Grassi, CEO, General Assembly
The message is clear: Without role-specific guardrails and upskilling, AI adoption doesn’t just stall—it becomes a risk.
What data professionals are using AI for—and what they want to learn
AI is already integrated into core analytics tasks. But there’s a clear desire to go deeper.
| Task | Already using AI | Want to learn |
| Analyzing data | 59% | 49% |
| Generating documentation | 49% | 33% |
| Generating ideas for analysis | 43% | 35% |
| Assessing data quality | 38% | 33% |
| Cleaning data | 36% | 31% |
| Writing code | 36% | 36% |
| Generating data visualizations | 26% | 35% |
| Developing predictive models | 19% | 32% |
| Debugging or pipeline troubleshooting | 20% | 25% |
What this shows
Even experienced data teams are still trying to understand how to apply AI to advanced modeling and automation, and they want clearer examples rooted in their day-to-day responsibilities.
The AI skills gap is real (and growing)
Despite widespread usage, confidence in AI skills lags:
- 63% feel very or completely confident in their AI skills
- 32% are only “somewhat confident”
- 5% lack confidence altogether
The confidence gap stems directly from uneven training access:
- Only 19% have received role-specific AI training
- 38% received training that was too conceptual or generic
- 25% received no formal AI training
- 18% had to self-teach
And the consequences show up in productivity.
AI is delivering benefits, but not for everyone
For many teams, AI is unlocking new levels of efficiency and creativity:
- 61% say AI frees up time for more strategic work
- 30% say it allows them to work fewer hours
- 85% say it improved company data quality
- 83% say it unlocked higher analytical creativity
- 79% say it enhanced organizational data literacy
Yet benefits aren’t evenly distributed:
- 18% say AI hasn’t impacted their daily work
- 19% say AI has actually created more work
This split is one of the strongest indicators that training, not tools, determines whether teams see meaningful gains.
Role-specific AI training drives better outcomes across the board
Data professionals who received job-specific AI training outperform their peers on nearly every metric.
They are more likely to:
- Use AI at work: 94% vs. 48% (no formal training)
- Complete planned daily tasks: 84% vs. 68%
- Feel confident in their AI skills: 91% vs. 62% (generic training) and 18% (no training)
And they report dramatically higher business impact:
- Data quality: 12% more likely to see improvements
- Team creativity: 13% more likely to see increased analytical creativity
- Data literacy gains: 18% more likely to report organizational uplift
- Better stakeholder communication: 8% more likely
- More time for strategic work: 8% more likely
The pattern is unmistakable:
When AI training is contextualized to the work people actually do, adoption rises and results multiply.
What type of training do data professionals want?
Data teams don’t want one-size-fits-all workshops.
They want training that reflects the complexity and nuance of their day-to-day workflows:
- 56% want self-paced online modules with data-specific examples
- 49% want updates as AI tools evolve
- 45% want interactive, use-case-driven workshops
- 42% want peer learning sessions
- 36% want ongoing support and troubleshooting
Today’s data teams need training that meets them where they are and grows with them, not generic overviews that become outdated overnight.
AI isn’t replacing data teams, it’s reshaping expectations
Despite common fears, most data team headcounts remain stable:
- 56% say they’re more productive with the same team size
- 26% say AI hasn’t impacted productivity or headcount
- 13% have added headcount
- Only 3% have reduced team size
But expectations around efficiency are shifting:
- 81% say leadership now wants teams to try using AI before requesting more help
This reality puts even more pressure on teams to adopt AI effectively—and safely.
Motivations, fears, and barriers to entry
Top data pro motivations for using AI:
- 26% want to improve personal productivity
- 20% want to improve work output
- 18% want to build their AI skills
- 12% enjoy experimenting with new tools
Top data pro fears for using AI:
- 38% worry AI could replace them
- 27% worry it hurts learning for entry-level talent
- 16% fear colleagues could be replaced
- 11% worry it makes work less enjoyable
And for data pros who aren’t using AI, the top barriers are:
- Not knowing which tools are useful (38%)
- No training (38%)
- Lack of confidence (27%)
- Privacy/compliance concerns (27%)
- Time pressure (13%)
So, what would help non AI users get started?
- Clear examples tied to daily tasks (51%)
- Company-approved tools (49%)
- Training on best practices (33%)
- Step-by-step guidance (31%)
- Hands-on practice (31%)
The TL;DR? Data professionals don’t just need access to AI. They need structured, role-specific pathways to adopt it confidently and securely.
For data pros and beyond: Role-specific AI upskilling is a business imperative
Across every metric—adoption, productivity, quality, creativity, literacy, and stakeholder communication—role-specific AI training outperforms generic training by a wide margin.
AI doesn’t unlock value on its own. People do, when they’re trained with the skills that match their roles and responsibilities. This is the moment for organizations to invest not just in AI tools, but in role-specific AI capability building that elevates entire teams and ensures safe, strategic adoption.
If you’re ready to empower your data teams with AI skills that drive real business impact, we’re here to help. Explore our enterprise AI training options today.