금융앱 내 VUI 송금 과정에서 사용자 발화 개입에 따른 경험 연구 : AI 비서의 개입 중심으로

국문 초록 (Abstract)

4차 산업 혁명의 시대가 시작되면서 AI(Artificial Intelligence)기술을 접목한 다양한 지능 정보 서비스들이 우리의 삶에서 차지하는 비중은 점차 늘어나고 있다. 그 중에는 인공지능(AI) 기술의 발달과 동시에 AI 음성 에이전트와 같은 과업형 에이전트에 대한 관심도 커지고 있다. AI가 널리 사용됨에 따라, AI와 효율적으로 상호작용하는 인터랙션에 관한 연구도 활발하게 진행되고 있는 상황이다. 하지만 기존 관련 연구가 AI의 소통 방식에 초점이 맞추어진 점, 특정 산업군과 직결되는 효과적인 활용 방법에 대한 연구는 다소 미비하다. 기존 금융 서비스에는 음성이라는 새로운 상호작용 방식이 도입되고 있으며, 이에 따라 기존 은행 업무는 사용자에게 적합한 VUI의 형태를 도출해 이와 같은 서비스를 제공할 필요가 있다. 따라서 특정 산업군에서 AI 비서가 과업형 도구로 활용되어 효과적으로 사용자와 상호작용해 최적의 사용자 경험 유도 방법을 도출하는 연구가 필요하다. 본 연구는 산업군을 금융으로 특정하고 금융 앱 내 송금 과정에서 사용자 발화가 개입할 수 있는 단계를 패턴화하여, 각 단계 에 따른 사용자 경험(보안성, 편의성, 유용성, 지속사용의도)을 연구했다. 이를 검증하기 위해, 실험자를 대상으로 인터뷰를 진행했다. 이는 2024년 6월부터 11월까지 약 6개월 간 진행되었으며, 20대부터 50대까지 총 72명이다. 이렇게 수집된 데이터는 Chronbach’s a를 통해 통계적으로 유의미하게 검증되었다. 주요 연구 문제는 첫째, “사용자 발화 개입 단계는 사용자 경험에 차이를 보일 것이다”로 이는 일원 분산분석(One-way ANOVA)을 통하여 사용자 경험 분석을 진행했다. 둘째, “사용자 발화 개입 단계는 AI의 개입 방법에 따라 사용자 경험에 차이를 보일 것이다”로 이는 이원분산분석(Two-way ANOVA)를 통하여 사용자 경험의 상호작용 효과를 분석했다. 결과적으로, ‘보안성’에서는 사용자 발화 개입 단계에 따라 모두 유의미한 차이를 보였다. 송금정보 입력・확인과 발화 없음 단계 모두 보안성에서 가장 높은 결과를 보였으며 이는 사용자 발화로 송금 발송과 직결되는 보안인증을 대체하지 않을 때 긍정적인 영향을 미치는 것으로 확인할 수 있었다. ‘편의성’에서는 사용자 발화 개입 단계에 따라 모두 유의미한 차이를 보였다. 모든 단계와 송금정보 입력・확인 단계가 가장 높은 결과를 보였으며 이는 사용자 발화로 인해 번거로운 수취인 및 계좌번호 입력을 대체할 시 긍정적인 영향을 미친다는 것을 의미한다. ‘유용성’에서는 사용자 발화 개입 단계에 따라 모두 유의미한 차이를 보였다. 송금정보 입력・확인과 모든 단계가 가장 높은 결과를 보였으며 이는 음성 송금 환경에서 사용자 발화 개입이 송금정보 입력・확인 단계를 포함할 때 긍정적인 영향을 미친다는 것을 의미한다. ‘지속사용의도‘에서는 사용자 발화 개입 단계에 따라 모두 유의미한 차이를 보였다. 발화 없음 단계와 송금정보 입력・확인 단계가 가장 높은 결과를 보였지만, 송금 결정・보안인증과 모든단계에서의 사용자 발화 개입은 부정적인 것으로 나타났다. 심층 인터뷰 결과, 이는 금융 서비스 특성상 비교적 보수적이며 철저한 보안인증을 경험해온 사용자가 기존의 방식을 고수하고자 하는 것을 의미하며, AI 음성 비서를 통해 과업을 진행하는 것에 대해 사용자의 신뢰를 느끼지 못하고 있기에 나타난 결과로 보인다. 마지막으로 사용자 발화 개입 단계와 AI의 개입 방법 간 보안성, 편의성, 유용성, 지속사용의도에서 상호작용 효과가 모두 유의미 하지 않다는 결과가 나타났다. 본 연구는 사용자 발화와 AI의 개입 방법의 적절한 균형을 찾아, 최적의 사용자 경험을 도출했다는 점에서 학술적 의의를 가진다. 현재 음성 금융 시장은 향후 5년 내에 두배 가량 성장할 것으로 예상되며, 이는 음성 인식 기술과 자연어 처리(NLP) 기술의 발전과 함께 개인 맞춤형 금융 서비스로 더욱 발전할 것으로 예상된다. 이에 본 연구가 음성 금융 서비스 사용자 경험 개선과 향후 발전에 기여하길 기대해본다.

목차 (Table of Contents)

• I. 서 론 ··········································································································· 1
• 1.1. 연구배경 및 목적 ························································································ 1
• 1.2. 연구절차 및 방법 ························································································ 3
• II. 이론적 고찰 ····························································································· 5
• 2.1. 음성비서 ······································································································· 5
• 2.1.1. 음성비서의 개요 ·················································································· 5
• 2.1.2. 음성 비서의 발전 ················································································ 6
• 2.1.3. 음성비서 시장의 현황 ········································································ 8
• 2.2. 금융과 AI ··································································································· 11
• 2.2.1. 금융산업과 AI의 결합 ······································································ 11
• 2.2.2. 해외 금융 AI 도입 현황 ·································································· 15
• 2.2.3. 국내 금융 AI 도입 현황 ·································································· 17
• 2.3. 사용자 경험 요소 ······················································································ 20
• 2.3.1. 보안성 ·································································································· 21
• 2.3.2. 편의성 ·································································································· 23
• 2.3.3. 유용성 ·································································································· 24
• 2.3.4. 지속사용의도 ······················································································ 25
• III. 연구 방법 ······························································································ 27
• 3.1. 음성 금융 비서 프로세스 연구 ······························································ 27
• 3.2. 연구 문제와 가설 ······················································································ 29
• 3.3. 실험 절차 및 대상 ···················································································· 32
• 3.4. 실험물 설계 ······························································································· 34
• 3.5. 설문지 ········································································································· 40
• IV. 연구 결과 ······························································································ 43
• 4.1. 인구통계 결과 ····························································································· 43
• 4.2. 측정변인의 타당성 검증 ············································································ 44
• 4.3. 가설 검증 ····································································································· 46
• V. 결론 ········································································································· 61
• 5.1. 결론 및 시사점 ··························································································· 61
• 5.2. 연구의 한계 ······························································································· 63
• VI. 참고문헌 ································································································ 66
• VII. 부 록 ···································································································· 76
• VIII. Abstract ······························································································ 85