The short story

• Challenge: Structurally protecting KBC’s retail customers against the threat of phishing.
• Solution: A phishing detection tool, a new service, integrated into the KBC mobile app. 
• Approach: Co-creating a tool with the KBC innovation team, tailoring the user experience to the nuances of the ML solution.‍
• Outcome: The KBC phishing tool, an in-app tool that identifies phishing messages.

The challenge

Phishing has become a huge societal problem. Fraudulent texts are a daily challenge to deal with. Scammers are becoming better at creating fake text messages that seem to be sent by trustworthy brands like ItsMe or Bpost. So it keeps getting more and more difficult to detect the malicious ones. 

As a pioneer in safe and secure financial services, KBC was looking for new ways to help protect its retail customers against the threat of phishing. As the responsibility still lies with the user, KBC wants to relieve this burden by helping to detect and prevent phishing attacks.

The solution

In order to protect users from fraudulent messages we designed and built a feature in the KBC app that helps detect phishing messages. In this phishing tool, all users of the KBC app can now easily check if a message is suspected to be malicious or not. Simply by uploading a screenshot of the suspicious text message (WhatsApp or SMS), the app can check what the chances are that the message is fraudulent. The application also provides advice about what to do when you receive a phishing message.

The approach

During the whole process, we worked closely together with Surf Studio and the KBC innovation and data team to create an application that is trained to detect, analyse & eliminate scams. AI-based solutions only work well if they support the user with a good user experience, set the right expectations and explain what to do when getting an alarming result. Thus, you have to mix ML with a good portion of UX.

Through the use of optical character recognition (OCR) the tool can retract a link from the uploaded screenshot and check where the link leads to. It will also analyse the use of words and tone of voice of the message since phishing messages often contain an element of urgency. Thanks to the collab with the data team at KBC, this twofold analysis leads to a colour code that indicates the possibility of the message being fraudulent.

The outcome

KBC's retail customers now have a powerful tool at their disposal to combat phishing scams, leading to increased security and protection against fraudulent messages. The integration of artificial intelligence and user experience has resulted in the accurate detection and prevention of phishing attacks, enhancing customer trust and confidence in KBC's cybersecurity measures.

Overall, the outcome of this tool has strengthened KBC's position as a pioneer in the financial service industry and demonstrated its commitment to safeguarding its customers' interests.