Since 2023, buy now pay later (BNPL) giant Klarna has been steadily replacing its human workers with more and more artificial intelligence (AI). It’s part of a cost reduction drive that is helping the firm prepare for its planned IPO and already, Klarna has reduced headcount by more than 1,000 workers while average revenue per employee has increased 73 percent when compared with 2023 figures.
But according to new research from ArvatoConnect, in the UK alone companies risk losing one in two vulnerable customers as the use of AI grows. According to the Financial Conduct Authority (FCA), 47 percent of individuals in the UK identify as having one or more characteristics that would classify them as vulnerable, including mental or physical health conditions, resilience difficulties, capability issues and trauma from bereavement and other life events.
ArvatoConnect polled 250 individuals who self-identify as vulnerable for its latest report AI, Digital Transformation, and Vulnerable Customers, which explores how businesses must blend human empathy with advanced technology to meet the unique needs of vulnerable customers.
As many as 78 percent of respondents said they look for “some level of human interaction” when seeking assistance. The vulnerable consumers said they have faced issues with the capacity of AI tools to fully grasp their individual needs, which causes them delay and frustration.A clear theme throughout the findings was that there is considerable room for improvement in how AI recognises and responds to individual vulnerabilities.
CX Network caught up with James Towner (pictured below), chief growth officer for ArvatoConnect, to find out more about “humanizing the digital and digitalizing the human”.
CX Network: The conversation around AI for CX is that AI should supplement the work of humans rather than replace them – particularly in the contact centre. What is your take on the news that Klarna could cut up to 3,000 workers as it prepares to list on the stock exchange?
James Towner: Advancements in AI have opened the door for huge efficiency savings in customer services, which Klarna has sucessfully embraced alongside OpenAI, the company behind ChatGPT.
But there’s still a limit to how far AI can help customers. Klarna’s pioneering approach to AI may be appetizing to potential investors, and for many customers too. But AI still struggles to match the emotional intelligence of humans that can detect the nuance of a person’s circumstance. One of consumers’ biggest customer service gripes continues to be not being able to get hold of a human.
Companies can get the balance right by adopting a hybrid approach that combines cutting-edge technology with a personal touch.
CX Network: Which customer journeys and situations do you believe AI is best suited to handle?
James Towner: As Klarna has shown, AI chatbots can provide significant value, especially for handling routine customer queries. However, it’s not about choosing one over the other. Empowering agents with AI tools can enhance service efficiency and quality.
Let AI handle the heavy lifting and analyze data to predict customer preferences, but ensure that real human support is always just a click away. This approach helps to create memorable "wow!" moments and make customers feel valued and understood.
By blending AI with human agents, we’re creating what we like to call “humanizing the digital and digitalizing the human.” It’s a hybrid approach where, for instance, live AI speech analytics can offer agents helpful prompts or real-time insights during customer conversations.
The evolution of these tools will continue to redefine excellent customer service. Advanced systems can analyze customer data, suggest personalized next steps and guide agents through sensitive interactions, enabling them to deliver faster and more tailored solutions to those who need it most.
CX Network: And which customer journeys and situations do you not believe AI is currently suitable to handle?
James Towner: Our recent study found how businesses risk losing a huge proportion of people that identify as having a vulnerable characteristic through an overreliance on AI.
We polled 250 individuals who self-identify as vulnerable and we found that more than two thirds (78 percent) still prefer some human contact when seeking help. Many felt AI struggles to fully grasp their unique needs, often leading to frustration and delays getting problems resolved.
Nearly half (48 percent) even said they avoid businesses entirely that provide inadequate support tailored to their needs – namely, a lack of access to a human agent.
That said, 56 percent of respondents to our study believe AI could meet their needs as well as a human, which shows growing openness to digital solutions, while reminding us that the want for human interaction remains.
The real takeaway here isn’t about choosing between AI or human agents - it’s about giving people options, to avoid running the risk of alienating customers.
CX Network: What detrimental impacts could AI-powered journeys in finance/ fintech, have on customers?
James Towner: Financial services customers often require more support than those in other sectors. Finance matters are deeply personal and high-stakes, involving sensitive data and complex decisions, which is why customers seek reassurance, clarity and tailored advice – elements that AI today will struggle to provide on its own.
While AI is efficient for handling straightforward queries, it lacks the empathy and understanding necessary for more nuanced financial issues. For complex scenarios like fraud, mortgage enquiries or large transactions, customers need to feel that their concerns are being addressed accurately and sensitively. AI’s nature to offer generic responses and overlook the subtleties of individual situations can lead to frustration and distrust.
In addition, customers with vulnerabilities – such as the elderly, those with disabilities or individuals facing financial hardship – are especially in need of personalized assistance.
Without access to human support, they may feel overwhelmed by technology, potentially leading to poor financial decisions or disengagement from financial services altogether.
CX Network: What checks and balances need to be in place to ensure AI tools are ethical and unlikely to harm customer outcomes?
James Towner: Transparency is key. Customers should always know when they’re interacting with AI rather than a human. Being upfront about this helps manage expectations and builds trust.
Regular audits are essential to ensure that AI systems operate effectively and fairly, helping to prevent biased or unjust decisions. Human oversight remains critical in this process. While AI should enhance the capabilities of human agents, it can also replace them in certain situations where tasks are straightforward and repetitive. For example, AI can draft emails and letters on its own, but a human agent should always review and finalize the correspondence before it is sent.
AI tools should also be tested across diverse user groups, including those identifying as vulnerable, to ensure they meet everyone's needs and prevent unfair exclusion. Customers should have accessible channels to provide feedback on their AI experiences, helping to identify shortcomings and guide improvements.
CX Network: As fintechs look to technology to automate more and more of what they do, what is your advice to prevent customer harm?
James Towner: As the Consumer Duty legislation enters its second year, driving a cultural shift in financial services toward positive customer care, firms are increasingly focusing on digital transformation to meet rising customer expectations and adapt to new technologies.
However, companies should avoid rushing into automation with a tech-first approach. It is essential to set clear goals first and ensure that customer satisfaction remains the key goal in all matters. Identify which processes need improvement, then design solutions specifically for those areas. It’s all about balance – automation can be a game-changer, but it needs to work alongside a people-first approach to avoid harming customers.
A great framework to follow is the ‘Four Ds’: define, design, develop and deploy.
This means that any transformation is built on a solid foundation and research-led insights, with a focus on enhancing the customer experience and not simply automating for its own sake.
Quick links
- Disrupting financial services: Fintechs, AI and true customer centricity
- How AI is changing physical retail
- Preserving the human touch in AI-powered CX
[inlinead-1]