JetBrains survey finds 90% of developers use AI coding agents weekly

Nine out of 10 professional developers surveyed now use artificial intelligence coding agents at work at least once a week, according to JetBrains' 2026 Developer Ecosystem Survey conducted between May and July. Nadia Rinsky, Head of GTM MENA at JetBrains, said the figure reflects how quickly AI is changing software development, with the focus now moving beyond code generation speed towards ensuring the software produced is safe, reliable and properly reviewed.

Human oversight becomes the priority#

Developers and companies must verify AI-generated code and accept accountability for the final software even as agents handle more of the coding workload, Rinsky said. Organisations now face decisions about which agents may touch their codebase, where proprietary information travels, and when human approval becomes mandatory before a change goes live. Project type determines the balance: prototype work may run with minimal human intervention, while production systems serving critical functions demand line-by-line human review of every modification the agent proposes.

JetBrains Air showcased at AI Everything Abu Dhabi#

JetBrains demonstrated JetBrains Air at AI Everything Abu Dhabi, a platform built to let development teams assign work to AI agents without losing human control over the output. Teams using the system can hand off coding assignments to agents, execute multiple assignments concurrently, and then examine, test, troubleshoot and approve what the agents produced. The platform integrates Claude Agent, Codex and Junie as available agents within the development environment, according to JetBrains. The company brought more than 30 development tools to the event, among them IntelliJ IDEA, PyCharm and the Kotlin programming language.

Cost and workflow management under review#

Organisations must now determine how AI expenses are tracked and how responsibilities split between human staff and automated agents, Rinsky said. Writing more code does not by itself improve engineering productivity, she said. The UAE's expanding use of AI in government and commercial operations creates openings to examine how autonomous agents fit into established workflows, Rinsky said.