First National Bank of Omaha (FNBO) — a $35 billion-asset, 5,000-employee privately held bank founded in 1857 and operating as the 15th-largest US credit card issuer and 4th-largest agribusiness lender — deployed Pindrop Fraud Assist, a GenAI-powered case intelligence add-on to Pindrop Protect, at its fraud investigation operations over a five-month rollout from August through December 2025. Within that window the bank's 4-5 person team recorded 35–40% faster case investigations (11–13 minutes saved per case), a 50% lift in analyst decision accuracy, and a 42% increase in case productivity without adding staff.
“When we went to executive management, their question was, 'How many additional heads are we going to need?' Our answer? None.” — Steve Furlong, Director of Fraud Management, FNBO. “Sometimes I had to listen to a whole 20-minute call to whittle it down to the couple of minutes that really matter.” — Steve Furlong. “Call summarization provides pretty quick hints on the direction of the case.” — Nate Koenig, Fraud Vendor Manager, FNBO.
Despite five years on Pindrop Protect with strong detection rates, FNBO's fraud investigation team was spending up to 30–35 minutes per case and relied on bilingual staff to handle language calls. Investigation decisions varied with analyst experience. The team wanted faster, more consistent case handling without increasing headcount or operational costs.
FNBO and Pindrop began their partnership over five years ago, building avoidance of millions in fraud losses on Pindrop Protect. In 2025 FNBO was among the first institutions to pilot Pindrop's first GenAI-powered case investigation tool, Fraud Assist, deploying it with the flip of a switch inside the existing Protect environment from August through December 2025.
Fraud Assist is a natively embedded GenAI add-on to Pindrop Protect that delivers real-time call summarization, multilingual-to-English translation, and auto-generated case notes. It required no new vendors, data pipelines, or infrastructure because it lives inside the existing Protect workflow, and analysts were able to bring Fraud Assist into daily operations almost immediately with minimal training.
The 4–5-person fraud investigation team handled 42% more cases by eliminating manual search and listening time. Analyst decision accuracy improved 50% with fewer false positives and false negatives. Non-bilingual team members reported less operational friction on language calls.