New Commonwealth Research Explores Intersection of Consumer Trust, Financial Services, and Conversational AI

BOSTON, Nov. 18, 2025 /PRNewswire/ -- New research from national nonprofit Commonwealth, explores how consumers with low-to-moderate incomes (LMI) perceive artificial intelligence (AI) in financial services and how their experiences impact their trust in the technology. As older, more limited customer support chatbots make way for advanced generative AI that can provide personalized financial support at scale, providers who understand how trust influences customer engagement can use these insights to design AI-powered tools that strengthen customer relationships and promote financial security. The report, made possible with the support of the Capital One Insights Center, builds on Commonwealth's earlier research. It can be downloaded here.

"Trust is the cornerstone of relationship-building. Conversational AI has the potential to dramatically enrich relationships with customers living on low-to-moderate incomes and contribute to their long-term financial health, but only if they trust it enough to engage and incorporate it into their financial routine," says Timothy Flacke, CEO, Commonwealth.

The research initiative featured a 3,000-person national survey exploring the experiences of U.S. households living on LMI ($30,000-$80,000) who use conversational AI chatbots offered by financial institutions; in-depth interviews; and field tests conducted in partnership with Metro Credit Union and Posh.ai, whose financial chatbot and voicebot banking tools were studied, and Arro. Commonwealth looked at five different aspects of trust—trust in financial institutions, trust in the effectiveness of AI, trust in data privacy and security, trust in oneself, and trust in providers—to identify challenges and opportunities for providers looking to build stronger relationships with their customers. 

Key findings:

• Customers expressed relatively high levels of trust in their banks, with 90% completely or somewhat trusting their primary bank, indicating it is not a barrier to engagement.
• Distrust of chatbots is driven more by concerns about information relevance and helpfulness than accuracy.
• A majority (54%) of users were concerned that the chatbot would either not understand them, would understand them but still not be able to provide an answer, or would simply take too much time or effort to engage.
• At the same time, 79% expressed trust in the information they received from their primary bank's chatbot, suggesting that the primary issue is not the accuracy of the information itself, but the ability of the chatbot to provide the information they actually want. 
• Human support matters. The option to transfer to a human agent is one of the top factors that support trust in conversational AI, preventing users from feeling trapped when they reach the limits of what a chabot can do.
• Data security and privacy concerns impact trust and engagement less than capability concerns. While data security and privacy do not necessarily prevent engagement with conversational AI, generative AI support may heighten these concerns because of a chatbot's ability to draw on personal and financial data as well as engaging in more detailed and personalized conversations.
• Generative AI presents a key opportunity to build trust by greatly expanding current conversational AI capabilities to provide personalized solutions. Trust in conversational AI's effectiveness remains limited primarily due to the limits of existing natural language processing (NLP) chatbots and their inability to provide truly personalized support. 

Building "Earned Trust"

Over the last 15 years, conversational AI—in the form of customer support chatbots—has become ubiquitous in financial services. More than three-quarters (76%) of survey respondents reported having used their bank's chatbot at least once. Among this group, about 25% said they use the chatbot weekly, while another 25% reported using it every few months. 

Conversational AI gives banks and other providers the power to create customized guidance tailored to individual customer requests at a scale and speed that would be impossible with traditional live customer support. Although this technology promises benefits for all customers, it is particularly important for households earning LMI. Commonwealth's prior research indicates that these households are nearly twice as likely to want to bank through personalized conversations, while at the same time have significantly lower rates of access to local bank branches. Personalized support through generative AI provides an opportunity to meet this unmet need at scale, building stronger customer relationships and improving financial health for this underserved population.

"Conversational AI goes beyond customer service and engagement," says Flacke. "This project uncovers insights that can yield financial benefits for customers, increase productivity within banks, and improve efficiency across financial institutions." 

Flacke adds, "The potential of generative AI makes it the most exciting technology to watch in the financial sector. Those who can earn trust through reliable generative AI tools will be at the forefront of a new era of deep, value-creating customer relationships."

For more information on Commonwealth's Financial AI work and this research, visit www.buildcommonwealth.org/our-work/financial-ai.

About Commonwealth

Commonwealth is a national nonprofit building financial security and opportunity for people earning low-to-moderate income through innovation and partnerships. Commonwealth collaborates with consumers, the financial services industry, employers, and policymakers. Because Black, Latin, and women-led households disproportionately experience financial insecurity, we focus especially on these populations. The solutions we build are grounded in real life, based on our deep understanding of people who are financially vulnerable and how businesses can best serve them. To learn more, visit us at www.buildcommonwealth.org. 

The generous support of the Capital One Insights Center made this report possible. All results, interpretations, and conclusions expressed are those of the authors alone and do not necessarily represent the views of Capital One, the Capital One Insights Center, or any of its affiliated companies.

Media Contact:
Jackie Jusko
216.374.0945
[email protected]

SOURCE Commonwealth