myCOI — illumend
Explore Project

AI compliance platform
Core Capacities
100+ enterprise customers
Project Focus
2025
Year
Overview
1
Transitioning from a static legacy system to a production AI platform — at enterprise scale, without downtime — required more than technical capability. It required a build partner who understood compliance deeply enough to codify it. The existing platform had no path to the real-time risk visibility or flexible rules processing that the market increasingly required, and the risk of getting the architecture wrong was high.


2
Sierra partnered with myCOI and Left Field Labs to co-design and build *illumend* from the ground up. The core engineering work centered on an application layer connecting document ingestion, AI processing, and compliance rule evaluation; a flexible rules engine adaptable to diverse client requirements and evolving regulatory standards; and forward-deployed engineers embedded directly with myCOI's team to maintain delivery velocity from concept to beta launch. The architecture was built for long-term maintainability — so myCOI could launch new capabilities without major rework.
3
• Translated myCOI's deep compliance knowledge into codified, automated rules processing at enterprise scale. • Built a flexible rules engine capable of adapting to diverse client policy structures and evolving regulatory requirements. • Integrated document processing technology for COI review accuracy and speed at high volume. • Forward-deployed engineers embedded with myCOI's team to maintain velocity and quality from concept to customer launch.

4
illumend has since scaled to more than 100 paying enterprise customers and reached mid-eight-figure annual recurring revenue. Clients report faster review cycles, improved visibility into compliance gaps, and meaningful reductions in delays caused by missing documentation. The architecture Sierra built gives myCOI a platform to expand into new verticals without rebuilding — and a competitive position that the legacy system could never have supported.