Evaluating AI Agents: Why It Matters and How We Do It

Abstract // As we integrate agentic AI into business products, robust evaluation of the agents is essential to delivering the highest quality. Proper evaluation ensures that AI agents are reliable, safe, effective, and aligned with user intent. Unlike traditional software or machine learning models, AI agents are non-deterministic and require specific types of evaluation. This talk outlines the importance of evaluating AI agents, the key components that we version and test at Acre Security, the metrics that matter for different types of agents, and how we currently achieve success evaluating AI agents that we build at Acre. Bio // Annie Condon// Annie Condon is an AI Solutions Engineer at Acre Security, where she helps bring intelligent systems to the physical access control space—without letting any rogue AI lock people out of a building (on purpose, anyway). With over 8 years of experience across machine learning, data science, and AI, Annie’s journey has taken her from deploying traditional ML models to building cutting-edge AI agents. Jeff Groom// Jeff Groom is an accomplished engineering leader specializing in AI-powered products. As the current Director of Engineering for AI-focused initiatives at Acre Security, he spearheads the development of domain-optimized solutions designed to enhance security across critical infrastructure sectors. With expertise that bridges advanced machine learning techniques and practical engineering execution, Jeff ensures that AI innovation directly aligns with operational and regulatory needs. Prior to his company being acquired by Acre Security, Jeff led engineering teams in the security space, driving architecture, development, deployment, and continuous improvement of AI systems tailored to real-world threat landscapes. Based in Denver, he is active in the AI and security technology community. An MLOps Community Production sponsored by Databricks