From content hoarding to data-driven decisions: Using AI to audit content at scale

Higher ed websites are often "digital junk drawers" filled with legacy pages, siloed content and significant content debt. While regular audits are the gold standard for site health, the reality is daunting: manual audits take months (if they get completed at all), organizational buy-in is tricky and the data is often obsolete before the audit is even finished. What if you could flip the script? In this session, we'll share a case study on how our team moved from manual audits to an AI-accelerated workflow. By developing a custom AI-driven auditing tool, we've been able to perform comprehensive WROT (weak, redundant, outdated, trivial) analyses on high-volume sites in days, not months. This talk isn't about hand-wavy, AI hype. We'll dive into the messy, imperfect reality. You'll learn how we structured our LLM prompts to recognize academic nuances, how we keep humans in the loop for accuracy, and how this shift allowed our strategists to stop counting pages and start improving them.