How Jit overcame developer resistance to shift to Cursor

Estimated reading time: 5 minutes

How the Israeli tech company overcame engineer pushback when it moved an entire team from JetBrains to Cursor.

AI-driven coding assistants have become the norm for the tech industry, with many companies implementing them into their operations, but are they just chasing the latest trends?

The security-as-code startup Jit implemented Cursor – the AI coding assistant that has attracted millions of developer fans – into its workflow to change its whole product delivery process. The adoption aimed to reduce manual and inefficient procedures from ideation through to production.  

However, Daniel Koch, Jit’s vice president of engineering, who spearheaded the transformation, made it clear to the team that the goal was not to implement AI for “the sake of saying [Jit is an] AI-native company”.

Instead, Koch aimed to “remove a lot of the manual friction [developers] have today.”

While inefficiencies became more noticeable to the engineering team, it was difficult to pinpoint precise bottlenecks without the right tools, Koch explains.

In turn, Jit replaced clunky steps with AI tools and prompts, using custom general-purpose technologies (GPTs) and model context protocols (MCP) in Cursor to update tickets or to address comments in pull requests. 

Koch found that Cursor can help developers write, understand, and fix code more efficiently.

Once he saw how quickly the engineering team could iterate and deliver results using Cursor, he “applied a similar approach to other teams, such as the product management team [and the] UX designers.” Eventually, it altered the entire process of how they built products. 

The engineering team was encouraged to “pinpoint every item in [their] process and ask: can we automate this with AI?” 

One key example of this is how Jit used the Cursor agent to reduce context switching issues many developers face. Part of this meant offloading tasks such as committing code, creating PRs, addressing PR comments, writing internal docs, creating tickets, and more to the AI agent, freeing up developer time for more complex tasks. 

“Our developers use Cursor for everything,” Koch says. 

Overcoming developer resistance 

However, with change always comes resistance.

In December 2024, the engineering team first started experimenting with Cursor, but only three out of 25 engineers were willing to use the AI agent. 

Many of the developers felt the code written by the AI agent resembled a “junior [engineer] – someone who just joined the team and didn’t understand [Jit’s] coding standards or conventions.”

However, Koch started to look into how Jit could configure Cursor to work with the engineering team rather than against.

“It wasn’t easy. Many developers had to move from the JetBrains IDE to Cursor, and it really is a different world. It’s not a simple adjustment. But I told them: we’re doing this together.

Koch was determined for the engineering team to learn how to use Cursor effectively rather than “roll back the change.”

In an effort to encourage more engineers to adopt Cursor, Jit organized “vibe coding” workshops daily from 12-14 March, which started with Cursor rules – structured instructions to help Cursor determine how and when to offer code suggestions, modifications, or automations based on the developer’s current context – and culminated in the creation of a fully developed product feature. 

“Setting clear Cursor rules is a crucial design decision that balances automation with developer control.” Koch explains, “Think of rules as instructions you will give a new onboarded team member on how we write code in a specific project – what development language is being used, which frameworks, what are our styling and code guidelines, or what is the project file structure. 

“It promotes safe experimentation and progressive rollout, reducing fear of unwanted or incorrect AI-driven changes.”

“Well-designed cursor rules become part of the ‘AI operating model’ inside the engineering org.”

Each session focused on practical use cases, along with a dedicated Slack channel that supported continued learning and collaboration.

“We showed that once we defined the right [Cursor] rules, we started getting better results. The agents could adapt, understand our codebase, and behave much more intelligently compared to working without any guidance.

“You can configure rules in Cursor’s settings to provide the agent mode all the context it needs to create code based on these instructions and act similarly to a team member,” Koch explains.

On the final workshop day, the team took a few options of end-to-end features, which included both front-end and back-end development, to “showcase the power of vibe coding.”

“That’s when something clicked. I could see it in their eyes – people were reimagining what was possible,’ Koch explains. “They realized this wasn’t just about generating a bunch of generic code. With the right setup, they could actually control and guide the AI effectively.”

“It was about learning how to truly integrate and benefit from AI.”

However, the challenges were far from over. 

“When we first started this shift, I told the team: I’m not sure everyone is going to love this. There may be people who won’t feel comfortable with the way we’re changing how we work – how we define our goals, how we measure success, and so on.”

This was proved to be correct as a couple months after the change was implemented, Jit had to part ways with one engineer who was not able to “adapt to the new way of working.”

However, Koch emphasizes that “the positive impact has been far greater than any of the downsides.”

Boosting engineering output

According to Stack Overflow’s 2024 annual survey of 65,000 developers, 58% noted improved efficiency, and 81% showed increased productivity with the use of AI tools. 

This was true for Jit as Koch acknowledges that just two months into the five month transition to Cursor the team already saw between 1.8x-3x productivity gains from ideation to production, measured against time to deliver new features.

“So we were able to shorten the time-to-value significantly – up to three times faster in some cases.”

But, it is not only Koch who has felt the benefits. “No one wants to return to the old way [Jit], worked before,” he says.

With Cursor now in place, Koch explains that the team revisited a long-delayed backlog of features that had been postponed due to shifting priorities and customer requests to see if they could approach them in a new, AI-assisted way.

Leading by example

Throughout the journey, Koch has learnt that he believes AI adoption must be intentional and collaborative. Leaders need to join their teams in using the tools, not just delegate the change. 

He also stresses the value of empowering excited team members as “agents of change” and warns that without grassroots involvement, top-down efforts will fall short. 

“It takes time and effort,” he says, “but the return on investment will be 10x more than the time invested in the shift.”

Get articles like this in your inbox

Choose your LeadDev newsletters to subscribe to.