• Design and test prompts: Work with engineers and PMs to shape AI behavior, define edge cases, and review outputs for accuracy and usefulness. • Own the ground truth: Build and maintain the “truth layer” — datasets and rating guides that represent correct, real-world GRC answers. • Evaluate and improve quality: Run side-by-side reviews, define launch-readiness criteria, and measure ongoing quality and drift after release. • Ensure responsible AI use: Help design AI systems that respect privacy, minimize hallucinations, and produce explainable, auditable results. • Document and teach: Write clear guides, checklists, and examples others can reuse; host short training sessions to raise the bar for AI quality across teams. • Collaborate widely: Partner with Product, Eng, and GTM teams to connect AI improvements directly to customer trust and business impact.