AI-Powered and Fraud-Proof
Customer Service for
a Large Credit Union

AI-Powered and Fraud-Proof
Customer Service for
a Large Credit Union


The No. 1 credit union mortgage lender in Colorado

A large Credit Union with $3B in assets,  serving 175,000 members, that provides a broad portfolio of financial products and services including checking and savings accounts, mortgage loans, credit cards, auto loans, home equity loans, student loans, business loans, and financial planning.


The customer was running an ageing Cisco on-premise infrastructure that they wanted to modernize and move to a fully cloud-managed Unified Communications (UCaaS) and Contact Center (CCaaS) architecture. They also wanted to add strong anti-fraud capabilities using biometric analysis of human speech and to enhance customer self-service using AI technology.


Omilia partner, RingCentral, implemented a combined UCaaS/CCaaS suite, and with our packaged ready-to-go integration Omilia deployed its no-code enterprise-scale Conversational AI platform, offering hyper-tuned out-of-the-box industry models.

Omilia’s Banking intent model classifies any human response to “How may I help you?” into hundreds of categories called “Intents”.  The simple, Wizard-driven configuration of the Intent miniApp allows the customer to deploy on their own in a matter of weeks without any prior experience using AI.


The power of Omilia’s Banking domain enabled the customer to provide fine-tuned routing to their members, ensuring they got to the correct representative the first time, using their naturally-spoken request. Transfers between agents were reduced substantially while Average Handle Time and Talk Time were also reduced since callers talked to the right agent on the first try.


Since go-live, 85 % of all members who call the customer’s main phone number and speak their request are classified in a matter of seconds. This has resulted in a dramatic improvement in caller experience, as members get what they need quickly and effortlessly.

What’s next

The customer is deploying Omilia’s Voice Biometrics capability to deliver Multifactor Identity Confidence as a deterrent to fraud, and also to reduce agent AHT further by authenticating members in the IVR using both a biometric analysis of their voice and ‘out of wallet’ information like ‘Date of Birth’ or ‘Account Number’.

Going forward, using the hundreds of possible intents identified out of the box, the company will provide endpoints in their data architecture that enables Omilia’s Conversational Voice service to provide answers to the most common questions. Omilia Conversational Voice will enable members to get simple, direct answers to simple, direct questions, creating an engaging experience that is quickly adopted by callers.

Over time, questions that were going to agents will seamlessly start being answered by the Conversational IVR.