Customer spotlight: How Adyen measurably improved developer experience
Kali Watkins
Product Marketing
We recently hosted a customer spotlight with Antoine Daignan and Micaela Stump from Adyen, where they shared how they improved their Developer Experience Index (DXI) by nine points in roughly a year, while simultaneously running multiple AI tool evaluations and rebuilding core parts of their release and local development infrastructure.
Adyen is a global fintech company with around 1,500 engineers, almost entirely on-prem, and an engineering org that has prioritized building product over internal tooling. This article summarizes what they focused on to improve developer productivity.
Where they focused improvements
DX gave Adyen a clearer picture of which problems were the biggest points of friction for developers. Early on, once they had implemented DX, two areas stood out: local development speed and release self-service.
Improving local dev setup
The environment setup was consistently taking more than 20 minutes. The team migrated from a script-based setup to Docker and used AI and scripting to auto-generate a ready-to-use starting environment for each team, reducing setup time to roughly 1 minute. The key to getting teams to actually adopt it was to do the work for them upfront. Instead of handing teams documentation and asking them to figure it out, they handed teams something that already worked. This resulted in 20% adoption in under two weeks.
Moving towards self-service releases
Adyen’s release process had been managed by a handful of operators since 2015, making them responsible for deploying thousands of applications company-wide. While it worked, it wasn’t scalable for the growing global climate that necessitates rapid iteration. The team rebuilt it as a self-service experience—app teams could now deploy their own applications through a structured rollout process, with built-in guardrails to catch issues early. Scripts and spreadsheets were replaced by a UI that let teams see and manage only their own apps. The result was that any app team could confidently own their own releases.
Addressing app ownership as a root cause
Low scores across collaboration, ease of release, and service catalog satisfaction pointed to a shared root cause: unclear application ownership. When ownership is ambiguous, no one invests in automations, documentation goes stale, and confidence in changes drops. Additionally, it hinders the impact of AI and future agentic use cases. The team worked hard to clean up ownership data, which led to a 22-point CSAT increase.
Making DevEx investments a team-level habit
Beyond the platform changes, Adyen invested in making DX something that individual teams could use on their own. Team leads were trained to run retros using survey data and to review dashboards with their engineers. The team published a DevEx newsletter, ran bi-annual showcases, and created a pilot volunteer program to keep teams engaged. More creatively, they put posters in the bathrooms, and saw more organic engagement from that than from most email campaigns. The sign that it worked was that engineering managers started pulling up DX on their own, without prompting.
How they’re approaching AI
Tool evaluation
Prior to April 2024, Adyen had no AI dev tools. Michaela and Antoine ran a structured evaluation of nine tools, assessing each against three criteria: use cases (code explanation, test generation, security), technical feasibility (monorepo support, model flexibility, compliance), and administrative factors (cost, support SLAs). Over 100 engineers participated across the POCs. Three to four tools made the cut. All contracts were kept to one year to stay flexible.
Thinking about token ROI
As AI usage grows, Adyen has introduced a monthly token cap per developer to keep costs in check, paired with training on using AI more efficiently. One example is using open-spec formats, which help models understand your codebase without having to reread the same context repeatedly. As they move forward, they plan to hold team leads accountable for their team’s AI spend, with the expectation that investment in these tools should translate into actual delivery outcomes, not just usage numbers.
Results and looking ahead
Since implementing these changes, Adyen has seen a nine-point increase in their DXI. Looking ahead, they plan to continue using data to identify where to invest, both broadly and by applying AI, and measure the impact of their developer productivity work.