When OpenAI recently announced that some ChatGPT users would be able to connect financial accounts directly to the platform, it marked another step towards a future where customers increasingly interact with financial services through AI-powered interfaces rather than traditional banking channels.
For retail banks, that development raises a broader question. If customers can increasingly ask an AI assistant to analyse spending, compare products, explain transactions or help make financial decisions, what role does the bank itself play in that experience?
Many retail banks work on the assumption that trust in customer support relies on human interaction, particularly when conversations involve sensitive financial matters. But new research suggests customer expectations may be evolving faster than many institutions recognise.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
Research across more than 1,000 consumers in the US and Europe found that people are three times more likely to prefer AI support for sensitive or personal conversations involving money or security than they are in general support scenarios. Overall, 61% of respondents either preferred AI support or expressed no preference when discussing sensitive issues. These findings don’t signal the end of human support – but rather they indicate something more nuanced: customers are becoming outcome-oriented in how they judge a service experience.
This research also reflects broader changes already taking place across financial services. McKinsey’s Global Banking Annual Review 2025 found that 23% of consumers are already using generative AI for financial tasks at least monthly, including understanding products, comparing options and receiving investment guidance.
The cost of friction
Consumers increasingly manage highly personal aspects of their lives through digital systems, including financial decision-making. Against that backdrop, discussing a repayment issue or suspicious transaction with an AI agent may not feel as unusual to customers as many banks assume.
In some cases, customers may value discretion and immediacy as much as interpersonal interaction, particularly in situations that feel stressful or embarrassing. The same research found that only 39% of consumers explicitly preferred humans when discussing sensitive issues, while 40% expressed no preference between AI and human support. More broadly, the strongest signal in the research was not necessarily enthusiasm for AI itself, but declining tolerance for friction.
While 68% of respondents initially stated a preference for human support, that shifted significantly once waiting times were introduced. Consumers became six times more likely to choose AI support when the alternative was waiting for a human response, and more than half said they would abandon a support queue within 15 minutes.
This presents a real challenge for retail banks, many of which still rely on fragmented support journeys, escalation queues and delayed resolution times. Customers increasingly compare banking support experiences not only against other banks, but against interactions they experience elsewhere in their lives.
In the near future, AI agents will increasingly become the interface through which consumers navigate financial products and services. For retail banks, that raises a broader strategic challenge: maintaining direct, trusted customer relationships in an environment where expectations around speed, convenience and conversational experiences are increasingly shaped elsewhere.
In that context, the competitive issue for banks will become whether customers continue tolerating operational friction that increasingly feels avoidable. Meanwhile, their competitors who have lent into AI are able to provide near frictionless experiences at scale for a large percentage of customer problems across chat, email, and voice.
What customers really want from their bank
One of the clearest themes running through the research is that consumers primarily want customer service teams, whether human or AI, to understand their issue. The most common reason respondents gave for preferring human support was not emotional connection, but the belief that humans better understand context and complexity. The forty percent who want “someone who understands my situation” are not voting against AI, they’re voting against being treated as a stranger.
As customer journeys become increasingly digital, expectations around continuity and context are changing. Consumers are becoming less tolerant of fragmented service experiences that require repeated handoffs or inconsistent answers.
This may ultimately reshape how banks think about customer support itself. The debate is becoming less about “human versus AI” in simplistic terms and more about which systems are capable of delivering accurate, contextual and efficient outcomes consistently at scale.
It turns out that AI is very good at handling data integration at scale whilst clearly understanding and acting on customer intent.
The next battleground is governance
While challenger banks are normalising AI-enabled customer experiences, many established retail banks remain cautious about large-scale deployment.
Consumer Duty obligations, vulnerability requirements and the EU AI Act’s evolving compliance requirements all place increasing emphasis on explainability, governance and customer outcomes. Any customer-facing AI deployment in banking will inevitably face heightened scrutiny around oversight, escalation and accountability.
However, regulation will ultimately shape how AI is implemented within customer support rather than prevent adoption altogether.
Many of the operational challenges regulators are concerned about – inconsistency, poor escalation, missed vulnerability signals and fragmented customer journeys – already exist within human-led support operations today. The issue banks need to contend with is badly deployed AI may simply add layers of complexity, but rolled out well, exponentially improves customer outcomes and operational consistency.
Institutions that move successfully in this direction are likely to be those that treat AI deployment as an operational and governance opportunity as much as a technology initiative.
The race to keep up
Retail banking has historically controlled the customer relationship tightly through channels, products and distribution. AI is beginning to challenge that model, and banks risk losing visibility into the moments where customer relationships are actually formed if customer expectations continue shifting faster than existing support and engagement models evolve.
In an increasingly AI-mediated environment, retail banks still retain one major advantage: a deep understanding of both their customers and the financial contexts shaping their decisions. Regardless of channel, the institutions best positioned for this next era are likely to be those that remain embedded in the customer experience itself – not simply as product providers, but as trusted, responsive and continuously present financial partners. If retail banks want to win, they’ll need to lean into this natural advantage alongside the new technology that AI unlocks to interact with their customers at scale.
Robbie Tilleard, GM EMEA, Lorikeet
