CHALLENGE: Why third-party insurance compliance still runs on manual review
The result is an industry where not understanding your own coverage is treated as the customer's problem, not the system's. Kristen Nunery, founder and CEO of illumend, knew that for too many businesses and individuals, third-party insurance compliance felt less like a system and more like a game of chance. “For most people managing COIs, it’s ‘trust your broker, you don’t know what you don’t know,’” said Kristen Nunery, founder and CEO of illumend. “And that gap in knowledge can lead to devastating consequences. I know—because I lived it.”
Nunery’s family lost their home in a fire and spent years fighting insurers for what they were owed. The complexity of their policy wasn't the problem. The problem was that no one had ever helped them see through it. That experience launched her mission to make sure no one else had to face risk blindly.
Managing certificates of insurance (COIs) has never been anyone’s idea of fun. It’s a tedious, manual process bogged down by legalese, contract scanning, and spreadsheet chaos. For businesses managing large networks of third-party partners, those delays create more than administrative frustration. Every missing endorsement, insufficient limit, or outdated certificate can introduce risk across contracts, projects and vendor relationships. The work is genuinely complex but for most people outside the insurance industry, it's been made to feel less like paperwork and more like decoding another language.
For decades, Kristen Nunery has worked at the forefront of COI management, bringing that deep industry expertise to illumend’s mission to modernize and simplify the process. As AI models matured to interpret complex documents, reason across requirements, and generate clear guidance, she saw the opportunity to scale and enlisted Left Field Labs to the modern era of insurance compliance software.
SOLUTION: How Left Field Labs built illumend's AI-first compliance platform
illumend is built as a system: rather than flagging coverage gaps for humans to chase, AI guides third-party partners through their own remediation. illumend uses a collection of models to analyze insurance requirements directly from contracts, verify policy documents, and provide fast, clear guidance. What once took humans up to 3 days, now takes 5-6 minutes, reducing review time by up to 99%. Left Field Labs engaged as a full-stack product and engineering partner, embedded in the business, present in daily decisions. We began with a structured AI assessment to define where the technology could have genuine impact, followed by a rapid prototyping sprint to prove out core technical concepts before a full build began. From there, Left Field Labs helped shape the product strategy, user experience, AI architecture, engineering roadmap, and production platform needed to turn illumend’s insurance expertise into scalable software.
Through this process, we helped to bring illumend to market as the world’s first AI-first platform designed to help businesses manage their third-party risk from intake through resolution.
The resulting platform combines four AI systems working in concert.
Agentic Assistant: At the center of the user experience is Lumie, a conversational AI guide that walks third-party partners through every step of the insurance compliance process, from document review and risk flagging to communication drafting and resolution.
Intelligent Document Processing: A document understanding engine processes thousands of policy documents daily, extracting structured data from unstructured, carrier-generated language.
Compliance Assessment: A compliance assessment layer, trained on each client’s specific requirements, evaluates every policy and identifies coverage gaps quickly and effectively.
Adaptive Exceptions (coming soon): A special circumstances engine that will handle exceptions: learning project-specific details to identify where requirements can reasonably be adjusted.
The architecture is built to evolve with the state of the art. Early builds required tight control-and-verification layers appropriate to GPT-3.5-era capabilities. As model reasoning improved, we shifted to broader instruction sets, leaning into the models' existing domain knowledge and building for future capability gains rather than engineering around current limitations.
The system is designed to adopt cutting-edge models as they emerge, ensuring illumend always runs on the best available performance. Just as important, the platform was built for trust in a high-stakes workflow. Full auditability runs throughout: every compliance decision, AI or human, is logged with its reasoning and rooted in source documents.
IMPACT: 99% faster COI review, 70% fewer admin hours, 90% AI accuracy
The result is a faster, more scalable, and more transparent third-party insurance compliance workflow that helps businesses reduce administrative burden while giving third-party partners clearer guidance on how to resolve coverage gaps.
Up to 99% reduction in COI review time, from as long as 3 days to 5-6 minutes
70% reduction in administrative hours required to manage COI compliance workflows
Thousands of agreements actively managed
90% average AI accuracy in compliance assessment, with documented cases where AI identified issues missed by human reviewers
First AI startup engineered and brought to market by Left Field Labs; with early market traction validating the product direction and competitors already imitating UI and core features
By combining illumend’s deep insurance compliance expertise with Left Field Labs’ AI product strategy and engineering capabilities, the team turned a historically-manual, expert-dependent workflow into an AI-first platform built for speed, scale and trust. For businesses managing third-party risk, illumend does more than identify compliance gaps. It helps partners understand them, fix them and move forward faster.