Regional U.S. Bank (unnamed)

A single compliance officer at a growing regional mortgage lender was manually reviewing thousands of ads per year via email chains, spending most of her time writing repetitive feedback about missing NMLS IDs, Reg Z disclosure gaps, and prohibited phrases. State examiner document pulls required days of hunting through shared drives and email folders, and loan officers viewed compliance as an adversarial bottleneck rather than a resource.

The bank deployed Luthor's AI-powered mortgage advertising review workflow, enabling loan officers and marketers to self-upload ads — PDFs, social images, email drafts, rate sheets, and co-branded builder pieces — for instant automated compliance checks. The system evaluated Reg Z trigger terms against required disclosures, NMLS and state licensing line presence, Equal Housing requirements, internal prohibited hot words, and RESPA co-marketing language, providing specific failure explanations and remediation guidance before ads reached the compliance officer.

Within a few months, ad review turnaround dropped by 82% and 86% of issues were self-corrected by the field before reaching compliance review. NMLS and licensing errors dropped to near zero, eliminating one of the most common examiner findings. State exam document pulls that previously took days of folder-hunting were reduced to minutes of export from a clean, metadata-tagged archive.

• When the AI tool (rather than a person) flags compliance issues, field staff accept corrections more readily and stop viewing compliance as adversarial — behavioral change follows when feedback feels objective.
• Role-based override controls are essential: giving the field autonomy to self-correct while reserving exception authority for compliance maintains regulatory integrity without creating bottlenecks.
• A structured, metadata-tagged archive of only approved consumer-facing ads — separate from drafts and internal comments — dramatically reduces examiner risk exposure and exam-pull labor.