Outdated Metrics are Starting to Crack

For years, NPS was the go-to measure of customer experience health. Ask customers how likely they are to recommend you, run the numbers, and you’d have a single, portable metric you could report upward with confidence. It was simple, consistent, and universally understood.

In 2022, Gartner predicted that NPS would begin a steep decline into irrelevance. NPS was built on assumptions about customer behavior that don’t really hold up in the real world. The idea that people regularly recommend businesses to friends and family in ways that are measurable and meaningful has never been convincingly demonstrated. At the point of a survey, people either don’t respond, or respond in ways that don’t accurately reflect their real behavior.

And this is reflected in the usage figures:

• Only 23% of organizations are still using NPS, in line with Gartner’s 2022 prediction that usage would fall below 25% by 2025,
• 80% of customer interactions are expected to involve AI in some form by the end of 2026, according to Gartner,
• Just 8% of organizations believe they’ve achieved a true, deep AI transformation, meaning most are still only scratching the surface.

The decline of NPS isn’t just about one flawed metric. It reflects a broader shift in how CX professionals think about measurement. Tools that try to infer customer sentiment through proxy questions and simple formulas are being replaced by solutions that can actually listen, interpret, and understand.

Sentiment Analysis and AQM are Transforming Processes

Here’s the fundamental problem with traditional CX measurement: it asks customers to tell you how they feel, rather than simply observing it. People are constantly signaling their emotional state through the words they choose, their tone of voice, and the issues they raise. That’s a rich stream of data that survey-based metrics almost entirely ignore.

AI-powered CX analytics taps directly into that stream. And nowhere is the impact more tangible than in Automated Quality Management, or AQM.

Before AQM, quality assessors could realistically listen to one or two calls per agent, per month. For large contact centers handling tens of thousands of interactions, that’s a vanishingly small sample. Most customer contacts were entirely unmonitored, which in the age of social media represents a very real reputational risk.

AQM changes the picture entirely:

• AI listens to 100% of customer interactions, giving contact center leaders a complete and accurate view of quality,
• Generative AI brings a qualitative dimension to the process, assessing tone, empathy, emotional context, and the themes customers are raising,
• Every agent gets individual, tailored coaching, with AI assessing performance at a granular level and identifying development needs specific to each person.

AQM certainly represents a step-change for CX Analytics, but it might be too early to write off standard quantitative metrics.

The Numbers Don’t Lie

It’s tempting, given the above, to conclude that traditional metrics are on their way out entirely. But that’s not quite the right takeaway. The decline of NPS isn’t a verdict on measurement itself, but a verdict on a specific kind of measurement that was always trying to do more than it was equipped to do.

Metrics like Average Handling Time and Customer Satisfaction score aren’t going anywhere. These “deterministic” metrics give CX leaders an accurate read on performance over time without trying to make claims about customer psychology. They report the facts of what happened, and that’s always going to be useful.

AI is valuable because it changes the context around the quantitative data:

• AI explains the “why” behind performance data, turning a dip in CSAT from a data point into an actionable story with identifiable causes and solutions,
• Qualitative proxy metrics like NPS are losing ground to AI-driven tools that can access genuine customer sentiment directly, rather than through indirect questioning,
• Deterministic metrics become more powerful alongside AI, because there’s intelligent interpretation to back them up.

For CX professionals, the goal in 2026 isn’t to choose between hard data and AI-powered insight. It’s to use both together; one as the foundation, one as an interpretive layer.

Building Trust in CX

Taking data security seriously isn’t just a regulatory requirement; there are consequences for Customer Experience as well. Customers take trust seriously. A business that looks after its data is likely going to look after its customers.

Data breaches, when they affect major companies, become global news. There’s no way to cover up mistakes; if your company suffers a data breach, your customers are going to know, and they’re going to take their business to competitors. In Customer Experience, trust matters. A single data breach can undermine a reputation built up carefully over decades.

Similarly, regulatory shifts can necessitate expensive restructuring. If you suddenly find your organization in violation of regulation, scrambling to adapt can be a costly and time-consuming process. By preparing now, you can save yourself time and money.

What matters is that you take control of your data strategy. It’s not enough to blindly trust in the cookie-cutter strategies of hyper-scalers. You need to understand your data, understand the regulatory environment, and work with a partner who’ll help you navigate uncertainty.

At Content Guru, we have two decades of experience designing resilient Customer Experience solutions for highly-regulated sectors. We’re the only cloud CX provider trusted by mission-critical emergency services. We’ll monitor regulations as they evolve, and ensure that your customer communications are always compliant. We always use in-nation data centers, ensuring that your customers’ information never crosses national borders.

Want to learn more about how the use of data in the contact center is changing? Discover our article on The Customer Data Platform and the Future of CX.