Verifizierung und Betrugsprävention

Krankenkassen: Schnellere Identifikation, schnellerer Service für Versicherte

Krankenkassen müssen sich stärker damit auseinandersetzen, wie sie ihre Prozesse umgestalten können, denn die Herausforderungen sind in den letzten Jahren gestiegen: Mitglieder erwarten heute auf vielen verschiedenen Kanälen schnell Hilfe und Beratung, die Datenschutzbestimmungen durch den Gesetzgeber werden immer schärfer und sie geraten immer öfter ins Visier von Betrüger*innen. Künstliche Intelligenz (KI) und Stimmbiometrie erhöhen nicht nur die Sicherheit, sondern verbessern auch die Customer Experience erheblich.

Wer bei seiner Krankenkasse seine Versichertenkarte beantragt oder Einblick in die eigene elektronische Patientenakte (ePA) erhalten möchte, muss sich dafür zunächst identifizieren. Früher erfolgte dies zum einen über die Versichertennummern. Da diese aber keinerlei persönlichen Informationen über die Versicherten beinhalten dürfen, wurden zusätzlich Antworten zu mehreren zufällig ausgewählten Sicherheitsfragen abgefragt. Dieser Vorgang nahm bereits viel Zeit in Anspruch.

Neue rechtliche Vorgaben erfordern nun im Onboarding eine zusätzliche Identifikation die entweder persönlich in der Geschäftsstelle der Versicherung, einer Postfiliale oder per Video-Ident-App auf dem Smartphone erfolgen muss. Geschäftsstellen und Postfilialen sind nicht immer direkt erreichbar und das Mobilfunknetz ist nicht überall so stabil, dass letzteres immer ohne Schwierigkeiten funktioniert, zumal diese Lösung Menschen ohne Smartphone ausschließt.

Die Identität wird unbemerkt verifiziert

Even later, the authentication process often does not get any easier when insured persons call their health insurance company. Regardless of whether they speak directly to a human or first end up with an Interactive Voice Response (IVR) system, they first have to verify their own identity. In order to simplify these processes from onboarding onwards, a voice bot can be used during the identification process to take over the data query. As the first point of contact, voice bots are able to resolve 85 percent of initial inquiries independently without having to involve agents. Overall, health insurers can thus increase customer satisfaction by 50 percent.

At the same time, newly insured persons will be given the opportunity to store their voiceprint in order to authenticate themselves via it in future calls. To do this, the voice of the insured person is analyzed within a few seconds during the call and compared with the imprint stored in the database on the basis of over a hundred characteristics. The best solutions, such as Nuance Gatekeeper, have a detection rate of 99 percent. In addition, the system is able to update stored voice profiles with new audio files, so that authentication works flawlessly at any time, even in the event of changing background noise or voice changes. Today, the use of biometric authentication and fraud prevention is already saving more than 2 billion euros worldwide.

Kriminelle haben das Nachsehen

With the help of voice biometrics, health insurance companies can also more easily detect fraudsters and prevent them from illegally obtaining benefits. The combination of insurance number and knowledge questions does not offer a high level of protection, as criminals can easily obtain this information, for example through social engineering attacks or on the dark web. With this data, fraudsters can claim to the insurance company that they have moved. They then request a new health insurance card, which is sent to the “new” address in order to claim services at the expense of the insurance company.

Bei der Authentifizierung anhand von Stimmbiometrie analysiert das System dagegen verschiedene Parameter wie Geschlecht oder Lebendigkeit der Stimme und kann auch Hinweise auf synthetische Sprache oder Sprachaufzeichnungen erkennen. Schlägt die Identifikation fehl, wird der Anruf automatisch für weitere Sicherheitsüberprüfungen registriert. Darüber hinaus können Krankenkassen auch die Stimmabdrücke von Betrüger*innen in der Datenbank speichern, sodass alle Anrufer*innen mit diesen verglichen werden. Somit können Betrüger*innen, die vielleicht gleichzeitig auch Mitglied bei der Krankenkasse sind anhand ihres realen Profils erkannt werden.

Health insurance companies must see their members more as customers who are willing to change providers if their service is unsatisfactory. Solutions such as voice bots and voice biometrics can improve the customer experience by enabling policyholders to voice their concerns and receive assistance more quickly, while maximizing the protection of their data. In order to gain access to all relevant information, the solutions must at least be integrated into the insurance software and the CRM system. However, the more health insurance companies want to tear down their data silos in order to expand their digital services, the deeper the integration must go. However, this can be achieved quickly without much effort via interfaces.