From frustration to delight: Designing the next generation of AI-powered banking chatbots

Chatbot adoption still has room to grow

According to Deloitte’s survey, 37% of respondents said they have never interacted with banking chatbots (figure 1). This suggests there is still potential for banks to increase awareness and usage of chatbots, possibly leading to cost savings and improved customer service.¹

Chatbot ratings reveal a generational divide

Deloitte’s survey respondents generally rate their experiences with banks’ chatbots positively. Millennials and Gen Zs, raised on digital interfaces, are more likely to report positive interactions, in contrast to baby boomers and Gen Xers, who rate their experiences less favorably. This finding highlights a generational gap (figure 2).

While these positive experiences are encouraging, it’s important to note that they primarily relate to simpler tasks. Deloitte’s survey reveals that users still largely prefer human interaction for routine banking queries.

Chatbots should focus on going beyond simple tasks

Banking customer respondents use chatbots mainly for technical support (60%) and inquiries about existing accounts (53%), as illustrated in figure 3.

As AI-driven chatbots become more advanced and personalized, banks can aim to expand the use of chatbots to revenue-generating activities such as responding to inquiries about new products and services, processing loan applications, and providing advice on investments and personal finances.² However, such a transition from primarily informational to transactional activities may be easier said than done.³ A separate study indicates that only a minority (27%) of its respondents trust AI for financial advice.⁴

The human touch is still important

Despite the prevalence of chatbots, most survey respondents still prefer to interact with human customer service representatives for simple, routine queries. According to Deloitte’s survey, 74% of respondents favored human agents over chatbots for these interactions (figure 4). Fewer respondents say they have used chatbots for more complex situations, such as personalized advice or applying for a loan, highlighting ways chatbots could evolve beyond basic functionalities.

Respondents who favored chatbots over human representatives did so primarily for the chatbots’ quick responses (78%) and ability to handle simple queries (69%).

This is consistent with previous research Deloitte conducted in 2021 during the COVID-19 pandemic, when many consumers preferred digital interactions. Among respondents who had used chatbots around the time the survey was conducted, many found them to be less knowledgeable and less empathetic than human agents.⁵ Of those who used chatbots for product inquiries in the past, 82% said they would not use them for this type of interaction in the future, and 46% indicated they would instead visit a branch. The resulting report considered that “humanizing chatbots … is going to be important in making digital banking behaviors stick.” This seems just as true now.

Accuracy tops the list of concerns related to chatbots

Among the respondents who preferred chatbots, 57% said that better accuracy was the improvement they’d like to see. Providing 100% accurate responses by AI chatbots could remain a challenge due to hallucinations, a problem that has been acknowledged for some time.

Also, nearly half of the respondents indicated that chatbots need improvement in areas such as enhanced personalization and stronger security and privacy (figure 5).

Lessons for the next generation of chatbots

As chatbots evolve into a new generation of sophisticated, agentic AI applications, consider how consumers interact and feel about them now. Addressing challenges—such as improving response accuracy, enhancing personalization, and ensuring greater reliability—will be important. Banks should also improve the precision and contextual relevance of responses and actions, and most importantly, design seamless collaboration between human agents and AI.

Getting from today’s limited, often frustrating chatbot interactions to more seamless, intelligent assistants won’t happen by accident. It should include deliberate investment in trust-building, smarter use cases, and deeply human-centered design. As banks chart this path forward, those that are successful will likely stop treating chatbots as mere cost-cutting tools and start reimagining them as intelligent partners in the customer’s financial journey, one interaction at a time.