FanDuel's Responsible Gaming platform supports player safety across the #1 gaming operator in the US, so every release has to be regression tested thoroughly across web and mobile. CreateFuture worked with FanDuel to build a self-healing AI testing agent, using the Claude API via Amazon Bedrock and Claude Code, that cut regression effort from an hour to six minutes per cycle and saves a full day of dedicated manual testing on every regression run.
regression effort per cycle, down from 1 hour
of dedicated manual testing saved per regression run
of the testing backlog now covered by the AI tool
FanDuel grew from its Edinburgh origins to become the #1 gaming operator and fastest-growing iGaming operator in the US. Acquired by Flutter Entertainment in 2018, it operates sports betting, online casino games, and fantasy sports across 37 US states.
CreateFuture has been a partner since 2020, with more than 100 experts working across FanDuel's business, from Safer Gambling and Player Management to Sportsbook and Casino.
Andy Giancamilli
Chief Operating Officer
Responsible Gaming is one of the most important parts of FanDuel's platform. It underpins the company's commitment to player safety and its obligations to regulators in every state it operates in, so changes have to be tested carefully before they reach players.
Regression testing on the platform took a day of dedicated manual effort for every run, and a large backlog of test scenarios was still waiting for coverage. FanDuel needed testing that could keep pace with delivery across web, iOS and Android, without taking engineers out of the loop on what changed in the codebase.
CreateFuture's engineers were already embedded in FanDuel's Responsible Gaming team, so they brought knowledge of the platform and its delivery pipeline alongside hands-on experience building agentic tooling with Claude.
FanDuel already describes expected platform behaviour as BDD (behaviour-driven development) scenarios. Rather than asking the team to write tests a new way, CreateFuture built an agent, powered by the Claude API through Amazon Bedrock, that converts those BDD scenarios into a reusable Playwright test suite using Playwright's MCP tooling.
When a test fails, a self-healing loop diagnoses the cause and attempts a fix. A FanDuel engineer reviews every proposed change before it is merged, so the agent does the repetitive diagnosis and repair work while FanDuel keeps control of the codebase. The team used Claude Code to build and extend the agent itself.
The agent runs inside FanDuel's existing BuildKite CI pipeline, so it fits into how the team already ships rather than sitting alongside it as a separate tool. It extends to native iOS and Android testing through BrowserStack's device matrix, giving coverage on real devices from the same BDD scenarios used for web.
regression effort per cycle, down from 1 hour
of dedicated manual testing saved per regression run
of the testing backlog now covered by the AI tool
Regression testing on the Responsible Gaming platform now takes six minutes per cycle instead of an hour, and the day of dedicated manual testing that each regression run used to need is no longer required. Half of the testing backlog is already covered by the new tool, across web and native mobile.
The agent builds on a partnership that has run since 2020, secured by a multi-year agreement in 2022, with some CreateFuture team members engaged with FanDuel for more than four years. It gives FanDuel a repeatable, AI-native approach to quality that its engineers stay in control of, and a model it can extend to more of the backlog and more of the platform.
Building a MoneySavingExpert app took brand and engineering excellence, and complex stakeholder management. But the result has helped over a million people save money.
![]()
Turning petabytes of data into useful information – and providing global access – was no mean feat.

Sprinting to a 40% conversion increase for online customer journeys.