Pritha Datta on turning fifteen years of banking expertise into live AI at Tredence
Where I Started
I joined Tredence about four years ago, carrying fifteen years of BFSI and data science experience with me. I knew the banking world intimately: its processes, its pressure points, and its deep resistance to change. What I didn't fully understand at the time was the potential that emerged when my domain knowledge combined with the AI talent that Tredence had quietly assembled.
The learning curve was steep and energizing. Tredence's culture pushed me toward technologies I hadn't worked with before, and the team around me — genuinely some of the most capable AI developers I've encountered — made it possible to move fast without cutting corners. Outside of work, when I'm not thinking about AI, I enjoy travelling and exploring new places. It gives me a new perspective to think about, adapt, and solve problems. Those perspectives have improved my patience and my openness to try new things.
The Moment That Mattered
My proudest win wasn't a prototype or a pilot. It was the moment an AI application for negative news screening went live and stayed live. Today, roughly 100 investigators use it as a standard part of their KYC onboarding process. It has delivered a 60% reduction in turnaround time for a workflow that used to be painstaking, manual, and chronically slow.
What made it meaningful wasn't just the number. It was knowing that the application had to survive the banking world's scrutiny — regulatory sensitivity, compliance requirements, and zero tolerance for unexplained outputs. Getting something that rigorous into daily production use, at that scale, is the kind of result I'll measure everything else against.
The gap between a good idea and a live banking application is enormous. Crossing it takes more than good AI; it takes people who understand what's actually at stake.
Who Made It Possible?
Two groups deserve real credit here, each in their own way. The senior leadership at Tredence gave me something I hadn't always had elsewhere: the room to experiment, get it wrong, and course-correct without the ground shifting beneath me. When clients pushed hard on timelines or scope, they stepped in — not to deflect, but to hold the line on what would actually produce a good outcome. That kind of backing changes what you're willing to attempt.
Then there's my team. The data scientists who worked alongside me on these projects didn't just execute — they problem-solved, stayed late, and cared about whether it worked in the real world, not just in a notebook. Watching them push their limits made it impossible not to push mine. That combination of supportive leadership and a team that drives results is what enables challenging projects to be completed.
What I'm Building Now
What's currently got me glued to my screens is an interesting use case in the AML space, where I'm working on a solution for a few banking institutions. There's a unique challenge in process management, aside from the technical hurdles. I'm excited to see the impact it will bring in turnaround times, because the automation is big on this one. What keeps me engaged is the same thing that always has: the banking world has no shortage of processes overdue for a rethink, and AI is finally sophisticated enough to do that rethinking responsibly.
Four years in, I'm more convinced than ever that the most valuable thing I can bring to this work is not the AI knowledge alone; it is knowing exactly where in a bank's operations that AI will face its hardest test. Over the next few years, I want to help build a team that combines domain expertise with cutting-edge AI engineering — because the real breakthroughs can come only from people who understand both worlds well.
Build at the Intersection. Join Tredence.
If the intersection of banking and AI is where you want to build, let's talk. Tredence works where deep domain expertise meets cutting-edge AI — in BFSI and beyond. If you want to build things that actually get used, by real people, in high-stakes environments, we're hiring.