AI Implementation blockers - The State of Data Series
What’s Really Blocking (AI) Progress in 2025?
In an industry that moves fast and talks faster, it’s easy to assume that the biggest obstacles to data and AI progress are abstract: fear of job loss, confusion about AI’s purpose, or a lack of executive buy-in.
But this year’s State of Data 2025 report paints a more grounded picture, one that’s both more practical and more solvable.
The top 3 blockers in 2025 are operational, not existential
When asked what’s really holding them back, UK data professionals didn’t cite hype or hesitation. They pointed to tangible operational challenges:
- Poor data quality (28%)
- Lack of in-house AI skills (21%)
- Managing stakeholder expectations (16%)
These are not philosophical barriers. They are process, skill and communication gaps that many high-performing organisations are already tackling head-on.
Why This Matters
The shift from cultural to operational blockers signals an industry that’s maturing. Previous blockers are being replaced by implementation challenges that require structure, not speculation.
This means that organisations aren’t asking “Should we do AI?” anymore.
They’re asking:
- “Do we have the right foundations?”
- “Are we ready to scale what we’ve already started?”
- “Who owns the outcomes?”
And that’s good news. Because unlike abstract fears, operational blockers are fixable, if you have the right talent, tools, and leadership.
Solving the Right Problems
Let’s break these blockers down:
Poor Data Quality
You can’t do AI, or even BI, on bad data. The most successful organisations are moving data quality upstream, embedding governance into ingestion pipelines and engineering standards.
Lack of In-House AI Skills
While external tools and vendors can help, real value comes from internal capabilities. Upskilling, cross-functional collaboration, and product ownership of AI use cases are key levers.
Stakeholder Expectation Management
As AI becomes more embedded, misunderstandings around capability vs. hype are creating friction. The best teams are communicating clearly, documenting limitations and aligning expectations from the start.
The Takeaway for Leaders
If your transformation efforts are stuck, it’s probably not because your team doesn’t care, or your tools aren’t good enough. It’s because your operations aren’t aligned with the realities of scaling data and AI.
Start by auditing:
Where quality control lives (and where it doesn’t)
Where AI skills sit (and where they’re needed)
Where expectations are misaligned (and how they’re managed)
Then, invest in change you can control and measure.
Download the full report for insights on how top-performing teams are solving these blockers in real-world environments.
Next in the blog series: What hybrid flexibility really means for data professionals in 2025.
