Omise AI
AI-powered efficiency
Back
Products
For Merchants
For Platforms and Banks
Professional Services
Resources
Help center
In the world of FinTech, innovation is often a buzzword. But here in Omise, we prefer to measure innovation in hours saved and prototypes built. Last December, we launched our AI Mastery Event Series: a high-octane journey designed to transform our engineering culture from “AI-curious” to “AI-Native”.
The series began with a session titled "Vibe Coding with Katti", delivered by one of our engineering leaders, Srinivas Kattimani. This Lunch & Learn session focused on a shift in the engineering mental model. Katti demonstrated how to leverage AI tools like Cursor to bypass the "manual friction" of traditional onboarding and complex architectural tasks.
Key technical takeaways included:
After kicking it off with the “theory”, the following week, we went with the “grind”. We took our existing backlog: 306 hours (over 38 person-days) of technical debt and localized bugs, and challenged our Dev and QA teams to tackle it using Cursor and GitHub Copilot.
The Result: In just 2 hours, the teams cleared 162 hours of work. That is a 52% reduction in the backlog at a speed. Our engineers proved that AI isn't just for writing new code; it’s a surgical tool for cleaning up the old.
An AI-native company doesn’t just build "cool tech"; it solves real business friction. Before the workshop ended, we invited our Product and Operations teams to pitch their deepest "pain points."
Our engineers moved from their IDEs to "Consultant Mode," interviewing internal stakeholders to identify where manual processes were slowing us down. This ensured that our upcoming Hackathon wasn't just a sandbox; it was a mission.
The series culminated in a 3-hour AI Hackathon. The goal moved beyond simple automation toward Agentic Workflows: systems that can reason, use tools, and make decisions.
The Standout Innovation:
Our winning team, Omise Compliance, addressed a massive bottleneck in our payment gateway service: Manual Merchant Reviews.
The Problem: KYC analysts were manually scraping merchant websites to verify if the business matched their application details.
The AI Solution: They built an Automated Merchant Review Agent. This agent autonomously scrapes a website, cross-references findings with application data, and utilizes AI reasoning to categorize the correct MCC (Merchant Category Code).
This isn't just a script; it's a digital workforce member that ensures compliance at scale.
While the tools (Bedrock, Claude, Copilot) were powerful, the real engine was the creativity and grit of our Omi-Squad. We saw senior mentors coding alongside junior devs, and facilitators acting as strategic architects.
We proved that when you combine human domain expertise with Agentic power, the "impossible" becomes a "3-hour sprint."
The AI Mastery series was just a taste. As we look toward 2026, Omise is committed to an AI-driven, agentic future. We aren't just "using AI", we are rebuilding our workflows around it. We are moving toward a state where every engineer is an architect of autonomous systems, and every manual pain point is an opportunity for innovation.
