Financial services has a CX problem. Regulators have noticed.

You receive a text from your bank about a fraudulent charge to your debit card. You call the number provided, navigate four menu levels, reach a chatbot that routes you to a human, then wait on hold for forty minutes, only to talk through your situation all over again.

This isn’t a hypothetical edge case. It’s the pattern Parloa’s State of Agentic CX report found by using our AI agents to assess 10,000 enterprise websites.  

At 64.2%, financial services and insurance companies represented the industries with the largest chatbot adoption rate. Yet, only 7.4% of those interactions successfully achieved the customer’s goal. 

When the chatbot was prompted to route to a human, the statistics weren’t much better. Chatbots delivered a successful handoff only 10.1% of the time.

In a similar vein, when exploring voice as a support channel, the research found IVR menu trees sending customers down six to eight levels of navigation before reaching the right path to resolution. When customers tried to talk to a human representative, they were often left waiting up to 90 minutes, with only 8% of businesses offering a callback option.

The results of this study suggest that while a majority of financial institutions are investing in automated CX, the areas in which they’re putting money are not bringing measurable returns. 

The architecture behind the numbers

Understanding why resolution rates are this low requires looking at what most “AI-powered” contact center systems really mean. Our global assessment found that of the 64.2% of financial services companies offering chatbot service solutions, 65.7% of them were offering rule-based systems.

Rule-based systems were designed to deflect. Companies think of the questions they believe customers are most likely to ask and build a one-way path to get the problems solved without human interference. The problem is, in today’s complicated financial landscape, customers are calling with questions unique to their specific situation. They’re seeking guidance in addition to transactions, and these personalized questions often break rigid menu trees.

Thegap between what customers need and what companies deliver has grown larger than any one-off situation. It’s become a global, regulatory problem. 

A global movement towards better CX

Around the world, countries are realizing the importance of positive customer experience and establishing regulation to ensure it for their citizens. 

In the UK, the Consumer Duty, which requires firms to demonstrably deliver good outcomes for retail customers, is central to the Financial Conduct Authority’s 2025–2030 strategy. It cites, “Ongoing consumer service should be as easy to access as sales, and firms should avoid barriers to switching, cancelling, or getting help.” With this regulation, brands cannot just say they deliver a good experience, they must use data to prove it. 

Based on our data, most companies would fail.

Under the FCA’s strategy, the repercussions of a poor customer experience strategy could be as light as the FCA instilling requirements to improve the experience or as heavy as withdrawal of authorization. In the UK, the risk of poor customer experience isn’t just losing one customer, it’s the risk of an entire brand’s existence.

Similarly, the United States’ Consumer Financial Protection Bureau (CFPB) has issued guidance on virtual agents in consumer finance, requiring not only accurate information but also guaranteed pathways to human agents. The CFPB encourages consumers to report incidents of chatbot failure.

Viewed through this lens, a 10.1% handoff success rate is not simply a technology underperformance metric but a measurable gap in the institution’s ability to follow the CFPB’s guidance, a gap eligible for reporting. 

From rule-based restriction to flexible automation

While countries are realizing the failure of rule-based CX systems to “solve” customer service, JD Power’s 2026 Retail Banking Satisfaction Study found that top-performing banks offering the least amount of service friction and delivering better problem resolution are best-positioned for continued growth.

The transformation to delivering such experiences requires a few key technological and organizational adjustments:

1. Institutions must stop looking at their contact centers as opportunities to deflect customers and instead focus on leveraging them as channels to connect with customers, strengthening customer relationships and increasing brand loyalty as a result.

2. Brands should consider building their support as an omnichannel experience, one where customers can call with questions and follow up via text.

3. Institutions must remove rigid menu trees and adopt an agentic AI approach to their service strategy, focusing on adapting to customers’ ever-changing needs and intent.

4. Brands should treat their support experience as one that is integrated with their broader customer experience, ensuring strong product integrations to keep all data up-to-date and systems compliant. 

There’s no time left to wait for a better customer experience

The most forward-thinking institutions are already reframing their contact centers not as costs to be contained but as relationship surfaces to be invested in. The landscape is changing fast, but those who move now may just be the ones to win over those who other brands lose. 

Customers who receive timely, effective help in a difficult moment tend to stay. More than that, they tend to trust. Agentic AI can not only help institutions deliver this level of support but also be adaptable for the future, when customer expectations inevitably change again. 

Download the State of Agentic CX Report to benchmark your business.