Transcription
U.S. Bank Case StudyBeth Gallagher DumkeVice President, Product Development & Innovation@GallagherBeth
U.S. BankU. S . B A NK 2
U.S. BankU. S . B A NK 3
Getting to PilotU. S . B A NK 4
Mobile Banking Platform Used a platform designed specifically for credit card access Integrated the NINA SDK into one app on the platform Designed the placement of the NINA dialogue box as well as themicrophone icon Out of scope was making any functional modifications to the app– Ultimately added transaction search functionality to the app as it was easyto perform with the NINA technologyU. S . B A NK 5
Use Case Definition & Grammars Worked alongside Nuance to review the existing app and determinewhere/how to integrate voice capabilities Ultimately landed on three use cases: navigation, bill pay, transactionsearch Nuance mapped out each interaction and intent that could beperformed within the app U.S. Bank team provided insight into the language used by customersfor grammar and language modelsU. S . B A NK 6
Pilot Scope Integrate the NINA SDK into our FlexPerks mobile banking appEmployees only12 week pilotBuild out three primary use cases:– Navigation– Bill Pay– Transaction Search Limit the navigation options to 7 intents to keep a narrow scopeU. S . B A NK 7
NINAU. S . B A NK 8
U. S . B A NK 9
Meet NINAU. S . B A NK 10
NavigationU. S . B A NK 11
Transaction SearchU. S . B A NK 12
Bill PayU. S . B A NK 13
Pilot GoalsU. S . B A NK 14
Learning Objectives Test the accuracy of using voice to navigate and perform tasks withinthe mobile app Understand if voice is a viable solution for navigation within the mobileenvironment Evaluate the user experience of using voice to interact with the mobileapp in place of tapping/swiping Understand the perceived benefits or drawbacks to using voice fromthe user perspectiveU. S . B A NK 15
Measurement Quantitative measurement was provided by Nuance Qualitative measurement was captured through a series of onlinesurveys– Surveys were sent weekly– Final survey sent at the end of the pilotU. S . B A NK 16
ResultsU. S . B A NK 17
Final Results NINA was invoked 45% of the time Users provided an average of 5.11 inputs per NINA session NINA had an overall recognition rate of 90%– Recognition rate improved as users became familiar with the capabilitiesDestinationsInputsTransaction SearchSpeechText/TouchFailureAccount SummaryBill PayUnsupportedCard BenefitsOtherU. S . B A NK 18
Qualitative Results 87% of users were satisfied with the app overall 78% of users said having voice enhanced their app experience 76% of users were satisfied with the ability to use their voice“I thought this wouldbe a great innovationand I was notdisappointed.”“It was able to distinguishwhat I wanted to do easilyin various noise conditions.It was so simple to use”“It was a lot moreresponsive than Iexpected. It had difficultywith a few tasks, butreally good overall”U. S . B A NK 19
User FeedbackU. S . B A NK 20
What’s NextU. S . B A NK 21
Moving Forward Results have been shared widely throughout the organization Looking for solutions that cross channels to provide customers acohesive customer experience Shortly after the NINA pilot, a pilot was conducted with Nuance VocalPassword (voice biometrics); the bank has put initial developmentfocus on biometrics Voice biometrics will be available to mobile banking customers laterthis year with a full roll-out in 2016U. S . B A NK 22
Questions?Beth Gallagher DumkeVice President, Product Development & Innovation@GallagherBethU. S . B A NK 23
U.S. BANK 22 Moving Forward Results have been shared widely throughout the organization Looking for solutions that cross channels to provide customers a cohesive customer experience Shortly after the NINA pilot, a pilot was conducted with Nuance Vocal Password (voice biometrics); the bank
