Scaling Transaction Monitoring Across Three Regions
Category
Compliance Operations
/
/
After building Rail's RFI process and the AI agent behind it, I had been running transaction monitoring largely on my own. When Rail was acquired by Ripple, the bar for transaction monitoring rose and a larger budget made a dedicated team possible. I moved from handling every review myself to building and managing the operation: training the team across regions, building a quality framework, tightening SLAs, and continuously improving the AI tool.
The challenge
I had built the RFI process from the ground up, but I was personally handling reviews 12 or more hours a day to cover Europe, the US, and APAC. This was during the acquisition period, when things were still settling and a lot rested on me to keep monitoring running, hence the hours. I was still improving things along the way, but it was improvement on the go, squeezed between reviews, never dedicated time set aside to do it properly. The acquisition raised the bar on transaction monitoring and, with a larger budget, finally made a dedicated team possible. The task was to scale the operation without losing quality or control. Concretely, a few problems had to be tackled together:"
One person spanning three regions, around the clock, to keep every market covered.
Reviews would now be delivered by an outsourced team, so quality had to be measured, maintained, and reported to management and regulators.
The team was new and remote, and needed to learn the transaction monitoring process to the same standard I had been applying.
Resolution times and the back-and-forth per ticket needed to come down, with SLAs that actually held.
The approach
I shifted from doing all the work to building the operation around me, in five parts.
Phase 1, build and train the team
I prepared the training materials and a knowledge base, then travelled to the Dominican Republic and Colombia to train the compliance teams in person, a week at each location. Once they were handling transaction monitoring on their own, I became their point of contact for escalations.
Phase 2, a quality framework
With reviews now delivered by an outsourced team, I built a framework to keep quality high and provable. I created a monthly sampling mechanism, tracked the results in a structured way, and produced monthly quality reports for each agent and for the team, the kind of evidence management and regulators expect.
Phase 3, SLAs, automation, and SOPs
I set rigorous SLAs and built a range of automations in Freshdesk to enforce them and take manual work off the team, from alerts when an RFI was close to breaching its SLA to routing, status updates, and other workflow triggers that kept tickets moving. Alongside them I wrote the SOPs, so the process stayed consistent and repeatable.
Phase 4, keep improving the AI tool
With dedicated time at last, every case and every quality review became feedback. I used them to measure not just the agents but how the AI tool was performing, and fine-tuned it continuously so it kept getting more accurate.
Phase 5, own the partner feedback loop
I managed the relationship with banking partners on RFI quality, confirming when their requests were valid, flagging mistakes, and pushing back on requests that did not make sense.
The impact
The operation went from one stretched person to a managed, multi-region team that improved month after month.
Coverage moved from a single person across three regions to a trained team, with me managing rather than executing every review.
Team quality rose by roughly 7% month over month, reaching 92% by the time I left.
Tickets in the queue were picked up and handled far faster, with time spent waiting for review down by about half, driven not only by the wider coverage but by the process improvements and queue-management automations behind it.
Around one fewer message per ticket on average, less back-and-forth and faster closure.
The AI tool kept improving, fine-tuned continuously from real cases and quality reviews.
Quality became measurable and reportable, with monthly per-agent and team reporting for management and regulators.
Scaling without losing control
Scaling a compliance operation is not just about adding people. It is about making quality measurable, repeatable, and provable. By training the team, building a quality framework, tightening the SLAs, and continuously improving the AI, the work went from one person stretched across three regions to a managed function that got better every month, with the reporting to back it up.
