The rise of intelligent Conversational AI has transformed how people interact with technology. Powered by breakthroughs in Deep Learning, Big Data, and Natural Language Processing (NLP), assistants such as Amazon Alexa, Apple’s Siri, and Google Assistant have achieved remarkable mainstream success. This progress has sparked growing interest in applying conversational technologies across various sectors, particularly in financial services. Conversational User Interfaces (CUI) are increasingly regarded as a promising pathway toward more natural and accessible human-computer interaction, with some researchers speculating that they could become the dominant universal user interface in the near future.

In the mobile banking sector, which has become the preferred channel for financial services, Conversational Virtual Assistants (CVAs) offer substantial benefits. Powered by AI, Speech Recognition, and NLP, tools like Erica (Bank of America), Eno (Capital One), and others deliver personalised insights, transaction support, account information, and 24/7 assistance. These assistants enhance convenience while enabling banks to reduce operational costs and improve customer engagement. However, many current implementations remain generic and lack sufficient personalisation.

A growing body of research emphasises that customisation and personalisation — particularly of personality, voice, language, and interaction style — significantly boost user satisfaction, emotional connection, trust, and adoption of CVAs. Despite this evidence, banking CVAs often rely on fixed, predefined settings with limited customisation options. This gap underscores the need for user-centred research to better understand diverse user needs and preferences, paving the way for more tailored and effective CVA design in mobile banking.

How can the CVAs in mobile banking be designed to create a more personalised and convenient user experience?

In line with the secondary research, Qualitative research findings highlighted the importance of CVA customisation and personalization. Quantitative data demonstrates that research participants’ preferences differ significantly based on the user personality type

The research findings were translated into an interactive CVA prototype developed in Adobe XD. This artefact effectively demonstrates the practical application of the user-centred design framework by visually communicating key customisation and personalisation features as well as multimodal user interaction flows. Adobe XD was chosen for its strong support of voice interactions, which aligns closely with the core focus of the project 

OpenAI GPT Agent Platform was used to demonstrate how the proposed design framework can be leveraged in practice to create a real-world CVA that could be integrated into the mobile banking app

 Future work should focus on conducting a usability study of the proposed prototype to identify and address potential usability or design issues. Further development is necessary to create additional interactions and user flows, as well as fully integrate this CVA into the mobile banking app. Also, the findings of this study should be validated in future research by engaging a broader population and increasing the sample size of participants.