4 Minutes
Developers aren’t resisting AI. They’re leaning in—just not blindly.
Across the board, there’s real enthusiasm for what artificial intelligence can do inside modern development workflows. Spotting issues before systems crash. Predicting failures. Tracing root causes in seconds instead of hours. The appeal is obvious, and the numbers back it up. But here’s the catch: excitement doesn’t equal trust.
Recent findings from Grafana Labs reveal a curious contradiction. While over 90% of developers see clear value in AI-powered diagnostics and forecasting, nearly all of them—an overwhelming 95%—want something first: an explanation. Not a result. Not an action. A reason.
That hesitation matters. Because when AI operates like a black box, even the most powerful tools start to feel unreliable.
Show Your Work—or Step Aside
Developers, by nature, are skeptics. It’s part of the job. When an AI suggests a fix or flags an issue, the immediate question isn’t “what,” but “why.” And right now, too many systems are failing that test.
Only a small fraction—around 15%—are comfortable with AI taking fully autonomous actions. The rest? They want transparency baked into every output. If AI can’t explain how it reached a conclusion, it risks being sidelined, no matter how accurate it is.
There’s also a practical frustration simmering underneath: context. Many developers report that feeding AI the right data, manually, eats into the very efficiency these tools promise. In other words, the setup cost sometimes outweighs the payoff.
Meanwhile, companies are juggling their own headaches—security concerns, compliance hurdles, and infrastructure that wasn’t built with AI in mind.
Europe’s AI Surge Comes With a Confidence Problem
Zoom out to the broader market, and the picture gets even more layered. AI adoption across Europe is accelerating fast. By early 2026, nearly 80% of businesses expect generative AI to be embedded in their workflows. Agentic AI—systems that can act independently—isn’t far behind.
On paper, it looks like a full-scale embrace.
In reality, there’s a gap. A wide one.
Research from Informatica highlights what it calls a “trust paradox.” Employees are using AI. They believe in its potential. They even trust the data behind it. But they often lack the skills to use it responsibly.
That disconnect shows up in training priorities. Data literacy—understanding the quality and meaning of data—ranks higher than AI literacy itself. And nearly all data leaders agree: their teams need more education before AI can be used safely at scale.
At the same time, governance is lagging. More than three-quarters of organizations admit their oversight hasn’t kept pace with how quickly employees are adopting AI tools. Instead of building tailored systems, many are opting for off-the-shelf AI agents, trading control for convenience.
It’s fast. But it’s messy.
Concerns are stacking up: data quality, security risks, lack of expertise—especially around autonomous AI—plus limited visibility into how these systems behave once deployed.
Speed Isn’t the Finish Line
There’s no sign of slowdown. Investment in AI is rising, with companies planning significant increases in spending across training, governance, and security. The ambition is clear: make AI not just powerful, but dependable.
The real shift now is from adoption to accountability.
Because rolling out AI tools quickly is no longer the benchmark. Trust is.
For AI to deliver meaningful returns, organizations need more than advanced models. They need clean, reliable data. Transparent systems. Teams that understand both the capabilities and the risks.
In the end, the companies that win won’t be the ones that deploy AI the fastest. They’ll be the ones that make it trustworthy enough for people to rely on without hesitation.
Leave a Comment