PayPal increases roadmap throughput by 40% with Cursor
PayPal rolled out Cursor across an engineering org of 8,000 developers and a decades-old codebase. High-adoption teams are now deploying daily and completed a 3,000-app Java upgrade in two months instead of a year.
PayPal has its startup energy again. Here's one measure of the change: PayPal recently completed a Java upgrade across 3,000 applications in two months. Previously it would have taken eight to twelve months. With 8,000 developers and a codebase spanning decades, that kind of acceleration seemed impossible.
But it's happening because of how the company adopted Cursor.
For the first time we're able to see a real path to being as scrappy as a startup, but with all the funding and maturity of a large enterprise.
Start with high-impact teams
PayPal recognized early on that AI needed to be an intrinsic part of its products and how it worked. But it also needed to move carefully. The company moves money for nearly half a billion people around the world, and the cost of mistakes is high.
So when they rolled out Cursor, they started with their highest-impact teams building the most impactful products where speed to market was essential. Within two weeks, those teams were deploying daily instead of weekly or biweekly.
Cursor has been critical in accelerating our timelines from doing something in four sprints to getting it done in one.
Adoption spread organically across the organization as engineers saw what their peers were accomplishing with Cursor. Teams with high Cursor adoption rates—many exceeding 90%—showed dramatic improvements in deployment frequency and lead time.
"Teams with high adoption are deploying faster and shipping with shorter lead times," says Chance. "Teams that haven't adopted yet are staying roughly the same."
New ways of working
The way PayPal builds software has fundamentally changed too. The old linear software development lifecycle—design, code, build, deploy—has given way to something more iterative. Teams move from idea to working prototype in hours, then iterate from there.
That process has collapsed traditional role boundaries. Product managers are bringing functioning prototypes to engineers instead of just PRDs, while engineers are replying with their own prototypes of what the product could look like.
Roles that used to be very finite are blurring and we're seeing better product ideas come out of it.
This shift, in process and in how roles are defined, is changing what matters in engineering. Software development used to be gated by experience with specific programming languages. Now it's much more about creativity and problem-solving.
Prakhar Mehrotra, Global Head of AI at PayPal, compares engineers using Cursor to the superhero Flash from DC Comics: a red-suited dynamo who moves beyond hypersonic speeds but still has to decide where to apply his powers.
Engineers still have to decide what problem to solve. But you're getting there much faster.
Measuring what matters
While the way PayPal builds software has changed, the company has continued measuring what has always mattered: deployment frequency, lead time from idea to production, change failure rate. They also deliberately avoid metrics that could distort how people work.
If you measure it, you impact it. If you tell a developer their success is based on what percentage of code was generated by AI, they'll just ask AI to write verbose functions.
Teams with high Cursor adoption are deploying multiple times per day instead of once per week. Lead times have shrunk. Change failure rates have dropped.
Based on those metrics, Chance recently told PayPal's leadership the company was ready to deliver 40% more capabilities in 2026 than it shipped in 2025. She's careful about how she frames that number, emphasizing that it's not about being 40% more efficient—a phrase that typically means doing the same work with fewer people.
I think that's really shortsighted. The real excitement comes from how much more we can do. When I see teams deploying faster and more often, that tells me this is real improvement and this is going to work.
The intelligence shift
Mehrotra sees what's happening at PayPal as the leading edge of a larger shift. The last 20 years were about organizing and distributing information: Google, mobile, Amazon. Something different is underway now.
AI is about intelligence, not information. It's a fundamentally different technology stack.
A fast, full embrace of that shift is allowing PayPal to transform its business, from payments to commerce more broadly. That expanded sense of possibility is pulling engineers in and making them want to contribute to what PayPal is doing.
The most important thing is our developers are participating in the AI journey. They want to come work for us because PayPal supports tools like Cursor.
If you're excited about leveraging agents to transform your engineering team's velocity, please reach out to our team to start a free Cursor trial.