NPS Still Reigns as CX’s Most Strategic Metric at HSBC

Stefan Witteman, Client Experience Lead at HSBC Private Bank, shares sharp insights on why NPS remains critical, why CSAT may be outdated, and how AI can elevate private banking—if guided by ethics, equity, and a human-first mindset.

NPS plays a critical role in HSBC Private Bank’s CX strategy. It’s a trusted, leadership-anchored metric that links directly to revenue and performance. Removing it would disrupt continuity and weaken internal CX momentum, says Stefan Witteman, Client Experience Lead at HSBC Bank.

“We only measure relationship NPS (rNPS) once a year as our clients are very busy and we don’t want to interrupt them too much with surveys. But it’s an important CX metric for the Private Bank as it clearly indicates relationship health, provides year-on-year trends, and correlates with core business outcomes such as revenue.”

Stefan helps improve the end-to-end experience for high-net-worth clients across global markets. With a background in client experience roles at HSBC, Barclays, ABN AMRO, Kindred, and Entain, he brings a unique blend of financial services and digital expertise. Stefan supports Relationship Managers and booking centres by driving journey improvements, measuring client satisfaction, and championing service enhancements in areas like payments and digital channels.

According to him, while CSAT has value in certain contexts, it is less strategic. It often lacks connection to financial outcomes and benchmarks, making it harder to justify investment compared to metrics like NPS that support long-term client relationship decisions.

He’s also optimistic about AI’s role in scaling personalisation but stresses the need for human oversight, ethical safeguards, and equity to avoid one-size-fits-all experiences or bias in financial services.

Excerpts from the interview:

If Net Promoter Score (NPS) were removed from your measurement framework tomorrow, what would the implications be—for better or worse? 

Removing NPS would be a big shift. It’s our main CX metric and appears first in our annual client relationship survey for a reason. We only measure relationship NPS (rNPS) once a year as our clients are very busy and we don’t want to interrupt them too much with surveys. But it’s an important CX metric for the Private Bank as it clearly indicates relationship health, provides year-on-year trends, and correlates with core business outcomes such as revenue. It’s embedded in leadership scorecards and widely used in performance tracking. It’s easy to understand, benchmark, and communicate across teams. Replacing it would require time to build trust in a new metric, establish historical context, and engage employees. I’ve also seen composite CX metrics being used before, but they often confuse more than they clarify from my experience. Ultimately, any core CX metric should correlate with financial value to support the business case for investment. So, while replacing NPS is possible, it would temporarily weaken our internal CX narrative and require careful transition planning.

ALSO READ: To Drive ROI, CX Leaders Must Move Beyond Vanity Metrics

In your view, are there any long-standing CX metrics that persist more out of tradition than impact? Which, if any, would you consider retiring—and why?

Customer Satisfaction (CSAT) is probably one I’d consider retiring but it all depends on the context. It typically measures single interactions, providing a limited view that rarely connects clearly to outcomes like retention or revenue growth. Without that business impact, it’s tough to build a compelling case for CX investment. Also, CSAT doesn’t offer much benchmarking value anymore. That said, if a metric drives real action in your organisation, it has huge value. The goal is to improve client experience and business outcomes. Not to obsess over the perfect metric. If CSAT gets people moving and makes sense for your business, then it works. But personally, I’d prioritise metrics that are easier to connect to financial value and long-term relationships. That’s where NPS can offer more strategic insight for long-term decisions. It also allows you to benchmark your performance against competitors.

How do you balance excitement and caution when considering the transformative impact of AI on customer expectations in banking?

AI is hugely exciting, but it’s important to focus on a few use cases that really deliver value. At HSBC, they took a measured approach. Starting small, forming a group of ambassadors, and collecting and driving practical use cases before scaling. There’s a big appetite for anything AI within the company and colleagues are eager to get their hands on it for everyday tasks like summarising documents, translations, etc. But more advanced use cases such as support with fraud detection or portfolio insights are also on the horizon. It’s important to give employees access to safe, approved tools and communicate clearly about what’s being developed. But to me the human element remains essential. I recently closed a long-held account with a digital wallet after a single frustrating chatbot experience. It simply didn’t allow me to speak to a human and couldn’t answer my question. This demonstrates how quickly poor AI implementation erodes trust. The key is to integrate AI in a way that enhances, not replaces, the human touch—especially in high-trust environments like private banking.

ALSO READ: AI in CX? Proceed with Caution (Red Flags Everywhere)

Do you think AI could inadvertently create a ‘one-size-fits-all’ approach in CX by narrowing down personalisation options to only what algorithms determine is best, potentially alienating diverse customer needs?

Yes, that’s a real risk. Especially if AI models aren’t properly governed. Algorithms can unintentionally exclude or favour certain segments. This we saw in the case of Amazon’s recruitment AI which had a gender bias. But when used responsibly, AI can actually unlock better personalisation and help businesses serve diverse needs more efficiently and effectively. Many companies struggle to tailor experiences for different segments. They are looking for bespoke experiences that are delivered persistently and consistently across every single interaction. AI has the potential to solve this by scaling personalisation in a smart and structured way. The challenge lies in designing systems that deliver consistent, fair, and differentiated experiences—guided by clear ethical principles and human oversight.

How do you address the potential for AI to create inequitable experiences? 

Ensuring equity in AI starts with responsibility. At HSBC, ethical AI principles guide how data is used to make decisions. Everyone from frontline staff to data scientists play a role in maintaining these standards. AI should be designed to respect privacy, protect data, and serve a clear purpose. Bias in training data is a major risk as it can lead to unfair treatment and damage trust. And this reputational risk is huge for an established bank like HSBC. That’s why explainability, proper controls, and regular audits are essential. Companies need to be transparent about how AI decisions are made and continuously improve based on learnings. Hidden biases or unintended consequences require vigilant monitoring and correction to make sure no one is unfairly treated. AI should elevate experiences for all clients and not just the majority that are easiest to serve or make the most money. 

ALSO READ: Most CX Dashboards Are Digital Graveyards