In Private, Secure, Conversational FinBots We Trust

TL;DR

This paper investigates how to design trustworthy conversational financial bots by studying user requirements for privacy and security, aiming to enhance user trust and support daily financial activities.

Contribution

It provides empirical insights into user preferences for privacy and security features in financial chatbots and offers design guidelines for trustworthy FinBots.

Findings

Privacy and security assurances are prioritized over conversational abilities.

Users desire transparent privacy policies and secure transaction features.

Designing for trust can transform FinBots into reliable daily financial assistants.

Abstract

In the past decade, the financial industry has experienced a technology revolution. While we witness a rapid introduction of conversational bots for financial services, there is a lack of understanding of conversational user interfaces (CUI) features in this domain. The finance industry also deals with highly sensitive information and monetary transactions, presenting a challenge for developers and financial providers. Through a study on how to design text-based conversational financial interfaces with N=410 participants, we outline user requirements of trustworthy CUI design for financial bots. We posit that, in the context of Finance, bot privacy and security assurances outweigh conversational capability and postulate implications of these findings. This work acts as a resource on how to design trustworthy FinBots and demonstrates how automated financial advisors can be transformed into trusted everyday devices, capable of supporting users' daily financial activities.