Contact Center AI

Besserer Kundenservice und weniger Betrug bei Finanzdienstleistern dank Conversational AI

Unterhaltungselektronik wie Smartphones oder Smart Speaker setzt Maßstäbe für natürliche Spracherkennung und -steuerung. Banken müssen damit erstmal Schritt halten. Einige stellen sich dieser Herausforderung bereits und nutzen Conversational AI. Damit wollen sie nicht nur ihre Kund*innen begeistern, sondern auch ihre Geschäftsziele erreichen. Dazu gehören beispielsweise eine höhere Effizienz in den Contact-Centern, aber auch ein besserer Schutz von Kund*innen und Mitarbeitenden vor Betrugsversuchen.

Conversational AI – the quality of AI that underpins a virtual assistant – is becoming increasingly important in everyday life today: In the last two years, for example, the prevalence of smart speakers is estimated to have almost tripled to 38 percent in the United Kingdom, and also more than half of Germans (52 percent) uses a voice application at least once a week. For established financial institutions, this results in a major challenge, but at the same time a unique opportunity.

The idea of settling financial matters through conversations with virtual assistants (VA) is far from new – but the real innovation is normalizing such experiences. Customers have become accustomed to chatting with Alexa or Google Assistant and have become familiar with apps like Cleo, where they can keep track of their spending in conversations with the virtual assistant. As a result, they now expect not only the ability to have natural conversations with virtual assistants, but also a smooth, personalized experience.

Established banks have to meet these high expectations. While many have self-service offerings, these are often based on a menu-driven IVR system and a simple chatbot, and will inevitably disappoint discerning digital natives. And since it is becoming easier and easier to switch providers, it is not only this target group that is more likely to look elsewhere: Studies show that the choice of bank and customer loyalty is most influenced by the “simplicity and convenience of service“.

While the challenge is great, the opportunity is all the greater: financial services organizations that respond to customer demand and create efficient experiences at the enterprise level using conversational AI not only optimize customer satisfaction, but also contact center effectiveness and fraud prevention. Some banks in the UK, Europe, and Australia are already successfully using Conversational AI.

Conversational AI bei der Commonwealth Bank of Australia

The Commonwealth Bank of Australia’s (CBA) virtual assistant “Ceba” won gold in the category “Innovative Use of Technology in Customer Service (Financial Services Industry)” at the APAC Stevie® Awards 2021. Its development shows what an initially quite simple VA is capable of when it is upgraded thanks to advanced machine learning and conversational AI.

Als Ceba eingeführt wurde, waren die Abfrage des Kontostands oder die Aktivierung von Bankkarten typischen Anfragen. Doch im Laufe der Zeit hat die Bank seine Fähigkeiten erweitert: Heute kann der VA etwa 70.000 verschiedene Fragen und mehr als 90 Prozent der Kund*innenanfragen in wenigen Sekunden verstehen und beantworten. Darüber hinaus kann er auch auf andere Self-Service-Optionen verweisen, Anfrage an Live-Agent*innen weiterleiten oder selbst bestimmte Aufgaben erledigen. Ceba erwies sich im Jahr 2020 als besonders wertvoll, als Fragen im Zusammenhang mit dem Coronavirus die Arbeitsbelastung massiv erhöhten – zeitweise um bis zu 1.000 Prozent. Allein im März 2020 hat der VA 581.000 Gespräche geführt, die sich in etwa vier von fünf Fällen um die Aussetzung oder den Aufschub von Zahlungen drehten. Die Flexibilität des VA war einer der Schlüssel zum Erfolg, denn sie ermöglichte es der Bank, die eingehenden Fragen im Blick zu haben und die Antworten fast in Echtzeit entsprechend zu aktualisieren.

Alles in allem ist Ceba ein großartiges Beispiel für das Potenzial von Conversational AI, die Kund*innenerfahrung und -zufriedenheit zu verbessern. So blieben die NPS-Werte für den Chat-Kanal der CBA im Jahr 2020 nicht nur konstant, sondern verbesserten sich sogar während einiger Hochphasen der Pandemie.

Conversational AI bei der Swedbank

Swedbank’s VA has also experienced an enormous rise. It started as an internal resource – a simple FAQ bot designed to help employees find information faster. Soon after, however, he was already supporting customers directly in the log-in area of the leading Swedish bank.

A key development step was the integration of Swedbank’s VA into the CRM system: this gives the VA insights into everything from the websites customers are trying to access to the cards associated with their accounts. As Hans Lindhom, Team Manager at Swedbank, explained to participants in a Nuance webinar: “By integrating with our CRM, customers can complete certain operations, such as replacing their card or viewing old bank statements, without having to contact an employee. It also helps us to give more individual answers.”

Die Automatisierung solcher alltäglichen Aufgaben kann erhebliche Auswirkungen auf die Effizienz eines Contact Center haben: Im Fall der Swedbank beantwortet der VA nun 80 Prozent der Kund*innenfragen und wickelt 25 Prozent aller Debitkarten-Wechsel ab.

Conversational AI bei BNP Paribas

Greater efficiency was one of the main objectives of BNP Paribas Personal Finance Spain when it launched its VA. “With our virtual assistant, we want to achieve the most important improvements of any customer service model: higher customer satisfaction and greater efficiency,” says Maria Paz Puchol, Director of Operational Digital Transformation. The decisive factor was to put together the right team. “The fact that members from different levels of the BNP Paribas Personal Finance team in Spain have been involved from the very beginning has contributed to the success,” explains Maria. “This made it easier and more precise for us to define what we needed.” The bank’s success proves it right – its VA has 96 percent accuracy in understanding requests, and the number of calls handled by contact center agents has dropped by 46 percent.

Jeder guter Bank-VA verfügt über eine explizite Strategie, um komplexere oder sensible Interaktionen an menschliche Mitarbeitende zu übergeben. Der VA bei BNP Paribas kann Kund*innen nahtlos an Contact-Center-Agenten*innen weiterleiten, wenn er feststellt, dass Fragen über seine Fähigkeiten hinausgehen oder die Kund*innen frustriert sind. Viele Finanzinstitute setzen KI aber nicht nur für sprachgesteuerte Bankgeschäfte ein, sondern auch, um die steigende Zahl von Betrugsversuchen zu bekämpfen.

Sicheres sprachgesteuertes Banking bei NatWest

In recent years, criminals have become increasingly adept at evading traditional authentication methods, whether by purchasing personal data on the dark web, stealing credentials and one-time passwords using malware, SIM swap attacks, or social engineering attacks on contact center agents.

Eine Lösung ist die Integration von KI-gestützter Biometrie in das sprachgesteuerte Banking und die Authentifizierung von Kund*innen auf Grundlage inhärenter Merkmale. Dies kann der Klang ihrer Stimme sein, aber auch die Art und Weise, wie sie ihr Gerät halten, darauf hin und her wischen und tippen. Stimmbiometrie-Technologie wird bei Banken immer beliebter, da sie den Kund*innenservice vereinfacht und beschleunigt und ihnen gleichzeitig ermöglicht, bekannte Betrüger*innen innerhalb von Sekunden zu identifizieren.

The NatWest Group also makes use of voice biometrics. Jason Costain, Head of Fraud, explains: “It’s not just about preventing financial losses, it’s also about stopping criminals. For example, we were able to identify a fraudster who was linked to suspicious logins to 1,500 bank accounts. This allowed us to protect potential fraud victims and identify ‘moles’ that the criminal network uses for their scams. This has led to two arrests so far.” In the first year NatWest deployed the biometrics solution, the bank reviewed more than 17 million calls, triggered 23,000 fraud alerts, and achieved a 300 percent ROI.

Conversational AI bei Finanzdienstleistern: Von den Vorreitern lernen

Finanzinstitute, die ihre Self-Service-Erfahrungen verbessern und mühelose Konversationen mit VAs ermöglichen wollen, sollten sich an den obigen Beispielen orientieren. Von nahtlosen Übergängen zu Live-Agent*innen bei BNP Paribas bis hin zur Überwachung eingehender Fragen bei CBA zeigen sie viele Best Practices, die für den Erfolg von Conversational AI im Bankwesen unerlässlich sind.

Sebastian Reeve

Über Sebastian Reeve

Seb Reeve is a customer experience industry leader who is always seeking to provide thought-leadership, lateral-thinking and decision-support for Fortune 1000 Enterprises who are both his customers and partners. Reeves has more than fifteen years of experience in deploying technologies to improve the user experience. In his current role at Nuance as EMEA Director of Product Management and Marketing, he is responsible for defining and evangelizing the Nuance customer care proposition across Europe, the Middle-East and Africa – sharing how companies can create extraordinary automated experiences which their customers actively choose to use rather than simply tolerate and complain about, promoting best practices in AI and Machine Learning to the world of Customer Experience.