There's a growing problem in customer service, and it's not what most executives think it is.
A new survey from ServiceNow and ThoughtLab just dropped some sobering numbers: 85% of customers prefer human support when dealing with complex issues. Meanwhile, only 1 in 10 executives plan to prioritize phone support over the next three years.
That's not a gap. That's a chasm.
The Numbers Don't Lie
The survey, which polled 1,350 customers and nearly 500 service representatives and executives, paints a stark picture of misaligned priorities:
- 59% of customers cite lack of empathy as their top frustration
- Only 25% of executives think empathy is an issue
- 50% of customers hate being transferred between departments
- Only 27% of executives recognize this as a problem
- 44% of customers are frustrated by repeating their issues
- Only 15% of executives acknowledge this pain point
Let that sink in. There's nearly a 3x gap between what customers experience and what leadership believes they experience.
The AI Paradox
Here's the irony: companies are racing to deploy AI-powered customer service tools to improve efficiency. And they're succeeding—at efficiency. "AI is really improving things from a speed point of view," notes Cristin Gooderham, ServiceNow's area vice-president.
But speed isn't what customers want when things get complicated.
The data shows customers are perfectly happy with chatbots and self-service for simple questions. Need to check your order status? Chatbot works great. Want to reset your password? Self-service is fine.
But the moment something complex comes up—a billing dispute, a product defect, a service failure—customers want to talk to a human. And not just any human. A human who actually understands their problem.
Why This Matters for Product Teams
If you're building products, this research should give you pause. Because the same empathy gap happening in customer service is likely happening in your product decisions.
Think about it:
- How often do you assume you know what users want?
- How much time do you spend actually listening to customer feedback?
- When did you last talk to a frustrated user—not to solve their problem, but to understand it?
The ServiceNow data reveals something product managers already know intuitively: there's a dangerous gap between what we think customers want and what they actually want.
The Hidden Cost of Getting It Wrong
The survey found that companies are already seeing high customer turnover from poor service experiences. And few organizations have made meaningful progress using AI to build emotional connections with customers.
"If we don't keep humans at the centre of those interactions, that's going to cause tremendous customer churn," warns Gooderham.
Churn is expensive. Acquiring a new customer costs 5-25x more than retaining an existing one. Every frustrated customer who leaves represents not just lost revenue, but wasted acquisition spend.
But here's what many teams miss: churn doesn't happen overnight. It's the result of a thousand small disappointments. A feature that doesn't quite work. Support that doesn't quite understand. Feedback that goes nowhere.
Bridging the Gap: What Actually Works
So how do you close the empathy gap? The answer isn't more surveys. It's not more data. It's better data—synthesized into actionable insights.
1. Listen at Scale, Understand Individually
The problem with traditional feedback collection is that it's either broad and shallow (surveys) or deep and narrow (user interviews). You need both breadth and depth.
This is where AI can actually help—not by replacing human connection, but by helping you understand the patterns in thousands of customer conversations. What are people actually saying? What emotions are they expressing? What's the underlying need behind the feature request?
2. Close the Feedback Loop
The ServiceNow study shows executives drastically underestimate customer frustrations. That's a feedback loop problem. The signal isn't getting through.
Product teams need systems that surface customer pain points directly to decision-makers. Not quarterly reports. Not filtered summaries. Real voice-of-customer data that makes the problem impossible to ignore.
3. Measure What Matters
Most companies measure satisfaction scores. But satisfaction is a lagging indicator. By the time your NPS drops, you've already lost the customers who would have told you why.
Instead, track leading indicators:
- Effort scores: How hard is it for customers to solve their problems?
- Sentiment trends: Is the emotional tone of feedback improving or declining?
- Feature-level satisfaction: Which parts of your product delight, and which frustrate?
4. Make Empathy Systematic
Empathy shouldn't depend on whether someone happens to read the right support ticket. It needs to be built into your process.
That means:
- Regular exposure to raw customer feedback for the whole team
- Friction logs that capture real user struggles
- Decision frameworks that weight customer impact alongside business metrics
The Path Forward
The ServiceNow research confirms what many of us suspected: companies are optimizing for the wrong things. They're building faster systems when customers want more understanding ones.
But this also represents an opportunity. The companies that figure out how to deliver both efficiency and empathy will have a massive competitive advantage.
The key is recognizing that AI isn't the enemy of empathy—it can be a tool for it. Not AI that replaces human judgment, but AI that amplifies human understanding. Systems that help you see patterns in feedback you'd otherwise miss. Tools that surface the voice of the customer before it becomes the voice of the churned customer.
The empathy gap is real. But it's closable—if you're willing to actually listen.
Building products customers actually want starts with understanding what they're actually telling you. Pelin helps product teams synthesize customer feedback at scale, turning scattered insights into clear priorities. Because the best product decisions start with the voice of the customer.