Gemini Code Assist Cut Their Dev Time by 37%. (ComplyAdvantage)

Financial crime doesn’t take breaks. ComplyAdvantage builds the AI-powered platform that financial institutions use to detect it in real time, processing sanctions updates within minutes of publication and scanning global news feeds for emerging risks. The platform handles the kind of work that used to require entire compliance teams, and it has to keep getting better as the threats evolve.

The problem was the codebase. Years of rapid growth had created the kind of technical debt that slows everything down. Original developers had moved on, leaving large sections of code undocumented. New team members and contractors had to lean on senior engineers just to understand what existing code was doing, which meant senior engineers weren’t building. When you’re racing to ship a major new offering on a fixed deadline, that overhead is a real problem.

What Gemini Code Assist Changed

ComplyAdvantage brought in Gemini Code Assist, starting with a pilot of 20 senior developers across three global development centers. The pilot ran methodically, with Jellyfish tracking development lifecycle metrics from Jira and GitLab so the team could compare results between Code Assist users and non-users.

The pilot group hit a 42% improvement in median development time. When ComplyAdvantage rolled out Code Assist to the full 170-person engineering group, the number settled at 37%. Commits per developer increased 50%. Merge requests per developer doubled. The major new offering shipped on time.

Support workflows changed too. The team processed more than 500 support requests through Gemini, automatically sorting them into genuine support issues versus bugs needing engineering attention. Accuracy in spot-checks came in at 75%. Root-cause identification that previously took hours now takes minutes. As Jennifer Bursack, Sr Director of Technology Operations, put it: “In one instance, a developer used Gemini Code Assist to identify the root cause of a customer’s bug within minutes. This previously would’ve taken hours of manual code investigation.”

Why This Story Matters

ComplyAdvantage is an ISV. It builds AI-powered financial crime detection software and sells it to banks and financial institutions. The Gemini Code Assist story isn’t about what their customers did with AI. It’s about what ComplyAdvantage did internally to ship faster, carry less technical debt, and get a 170-person engineering team producing more with the same headcount.

That’s a pattern worth paying attention to. Developer productivity tools like Gemini Code Assist don’t just save time. They change what’s possible to ship in a given quarter. For a company in a competitive, fast-moving space like financial crime prevention, that difference is meaningful.

Want to go deeper?