Building a great AI assistant is only half the battle – knowing whether it’s actually great is the other half Our team owns the measurement and quality layer that make Glean’s Assistant and Agents reliably better over time: evaluation pipelines, quality eval-sets, LLM-powered judges, agent observabi