Almost all developers want AI help, but it has to explain itself first
95% of developers say AI should explain its reasoning so they can validate responsesThey're worried that manually explaining context is eliminating time savingsAI could unlock observability data, but silos need to be dealt with first
New research from Grafana Labs suggests that developers are highly keen to adopt artificial intelligence, with 92% seeing its value in detecting issues before downtime and 91% appreciating it for forecasting and root cause analysis.
Although the benefits are clear, developers still lack trust when it comes to using AI – 15% don't trust autonomous actions, and an overwhelming majority (95%) prefer AI to explain its reasoning so that they can understand how it reached its output.
And while companies are struggling with security, compliance and unsuitable infrastructure, developers themselves are most worried about having to manually provide context, which takes away from AI's efficiency promises.
Article continues below
Developers are ready for when AI can actually boost productivity
"Practitioners want trustworthy AI that reduces toil and helps them move faster," Emerging Products VP Marc Chipouras wrote.
Setting out the roadmap ahead, Grafana Labs explains that developers want AI assistance and transparency to mature before autonomy becomes the default.
Elsewhere, the report also highlights that companies are struggling with observability. Three-quarters (77%) of those with centralized observability claim to have saved time or money, however many are still struggling with complex, siloed setups.
It's this observability that's set to get a spending boost in the next year, with leaders expecting more efficient operations.
The report also indicates how AI could help in this area, with a third (30%) of workers citing alert fatigue as the biggest obstacle to faster incident response.
Looking ahead, it's clear these two business priorities are closely intertwined. With effective observability generating huge volumes of data, an AI layer on top of that can make it easier for workers to identify patterns and reduce downtime.
