Software development is undergoing a major change in the way developers work with and create code.  

AI in software development has moved beyond the novelty of “look what it can generate” and into something even more useful: agentic engineering. 

That shift doesn’t remove developers from the equation, it changes what great development looks like. Instead of hand-crafting every line, teams are increasingly designing repeatable, governed workflows where agents can investigate, propose, test and action work across the SDLC (software development lifecycle). That’s where the real shift comes in. AI doesn’t just write code, it helps you to run all of your work around code. 

From vibe coding to agentic engineering 

From the onset of AI, the term “vibe coding” was often thrown around, referencing a genre of coding where developers were experimenting with AI, inputting prompts and seeing what comes out: prompt, generate, tweak, repeat. It was exploratory, playful and helped developers to build an understanding of what AI could and couldn’t do. 

While vibe coding still exists, from a professional developer standpoint, the preference is for agentic engineering. The act of working with AI has evolved to put the onus on managing how the AI is created to support the development lifecycle. This gives the developer more control over the outputs and the process to deliver higher quality results, more consistently. 

Repeatable patterns. Manageable controls. Measurable quality. 

Why agentic engineering matters 

Here’s what many development teams have already figured out, which is driving the move towards agentic engineering: the biggest time drain in delivering software often isn’t typing. It’s everything around it – investigation, coordination, testing, verification, documentation and endless context switching. 

Agentic engineering targets these friction points, giving you agents that take on well-defined tasks across tools and environments. The agents target those tasks, while you take responsibility of the decisions that make a real difference. 

JetBrains Central – the platform for agentic engineering 

As teams move from isolated assistants to multi-agent workflows, governance becomes the make-or-break factor. The more autonomy you introduce, the more you need visibility, control and shared context. 

That's the thinking behind the release of JetBrains Central – a unified platform for all you AI tools. Central offers a governed execution layer for autonomous agents. 

In Central, you can trigger multi-agent workflows with shared context and intent, embed your agents and gain the visibility and control to make agentic engineering possible. By reframing software development into a distributed system of developers, agents, tools, and infrastructure, Central gives you the oversight you need to effectively build with AI.  

Junie – beyond code generation 

If Central is the governed layer, Junie is the “do the work” layer. JetBrains Junie does more than simply write code. From handling testing and verification to debugging and maintenance and even supporting with documentation and knowledge sharing. 

Junie supports the SDLC end-to-end, from planning, to implementation, right through to maintenance and integration with IDEs and CI/CD pipelines. This enables you to focus on the intent, architecture and decisions. It’s the perfect tool for developers looking to use AI, not just to write code, but to enhance the way they work, across tools and platforms. Junie gives you the keys to manage and deploy agents to fit your needs. 

AI Assistant – transform your IDE into a supervised agent workspace 

Another reason developers are switching to agentic engineering over vibe coding is the productivity benefits it offers. AI Assistant by JetBrains is built with you in mind, to optimise your workflow. 

AI Assistant gives your agents real context, so they can ground their decisions in your actual codebase. Your files, symbols, recent changes and project structure are all capable of being reasoned over so your agents can act coherently across your system, keeping you in the flow of work while using AI.  

It’s also where agentic engineering becomes tangible day-to-day – delegating multi-step workflows when tasks are complex, and handling the smaller admin work that quietly drains momentum. 

JetBrains’ new direction: open, governed agentic development 

The AI landscape is moving fast and the industry shift from coding companions to agentic engineering is accelerating it. Instead of asking AI to generate isolated snippets, teams are starting to delegate end-to-end development work to agents that can operate across tools and environments under clear human oversight.  

In response, JetBrains is leading a shift toward a new direction: the JetBrains AI ecosystem is no longer just about faster code generation - it’s about agentic added value: UX and features that become available as and when they’re needed. 

JetBrains are now focusing on building an IDE experience that doesn’t depend on a single vendor or tool. Bring your own API key, stick to JetBrains AI-managed set up, connect external coding agents – the choice is yours. Whatever you choose, the JetBrains IDE experience is built to support you to effectively work with, build or manage agents.  

Which AI tools are right for me 

Wondering which JetBrains AI tools are right for you, or how JetBrains Central could help you govern and manage your AI agents? Get in touch. Our team can help with any questions you may have and get you set up with the right licences to take the next step.