The best customer insight comes from knowing what to listen to

One customer complains about a service, while another praises it. Most people stay silent, but their choices — what they search for, buy, abandon, return, or choose again — still show what they like. Today’s consumers rarely stop giving feedback, even without speaking. The question is whether companies know how to listen.

For decades, businesses created specific moments to understand consumers. They did surveys, ran focus groups, conducted interviews, and hired consultants to find out what people wanted.

Today, that feedback arrives constantly. Reviews capture reactions after a purchase. Customer service chats highlight problems as they happen. Searches show interest, while returns can signal disappointment without a customer explaining why.

FEEDBACK NEVER STOPS

AI has made it possible to analyze huge amounts of this information in real time. That gives companies something traditional research that captures attitudes at a particular moment can’t really provide: a continuous view of customer behavior.

Siddharth Sudhakar, General Manager, MENA at Trip.com, sees real-time feedback as a way to fill that gap. “Traditional market research remains useful, but real-time feedback gives us a much more continuous view of how traveler preferences are evolving,” he says.

Trip.com can look beyond what customers tell the company directly. Searches, browsing, bookings, customer service interactions, and post-trip reviews all provide different signals about what travelers want. 

John Bevan, CEO of dnata Travel Group, also sees direct feedback as more useful for answering a specific question: Is the company getting things right for its customers?

Dnata surveys customers after they travel. Some of its tour-operated businesses go further, calling customers after their return to ask what worked and what needs improving. The company uses those responses to shape future holidays and assess suppliers. 

“You can’t just do a survey and ignore it,” Bevan says. 

He adds, “That’s all about the personalization. You can’t do personalization unless you listen to your customers. So the only way you can get the feedback is by asking directly.”

In some cases, feedback has led Dnata to stop selling particular suppliers after repeated negative responses and little indication that the problems would be fixed. 

AI FINDS THE PATTERNS

However, the problem with continuous feedback is volume. A large consumer business can collect reviews, searches, purchases, cancellations, chats and other interactions at the same time. The information may be valuable, but only if the company can connect it.

This is where AI changes the equation. At travel platforms, for instance, AI helps analyze signals across different stages of a customer’s journey. Search behavior can be viewed alongside booking patterns, reviews, and customer service interactions to identify changes in demand. 

“AI allows us to connect signals that would be difficult to analyze manually,” Sudhakar says. 

He adds, “We can look across search behavior, booking patterns, reviews, customer service interactions, and other points of the customer journey to identify changes in demand, emerging destinations, shifts in booking windows, or changing expectations around price, flexibility, and experiences.”

For example, more searches for a destination matter even more if booking times are also shifting or a certain type of accommodation is getting popular. Looking at these signals together can reveal changes sooner than looking at each one alone.

Customer behavior can also show companies where customers are struggling, even if they don’t complain directly. Repeated searches, abandoned journeys or returning to the same information can point to friction in the customer experience. These signals can help identify where a product or service needs to become simpler.

Sudhakar says travelers often move between destinations, flights, hotels and activities before making a decision. 

The company has used those behavioral insights in developing tools such as TripGenie and Trip.Planner, designed to make travel planning more personalized. Dnata is also using AI to process large datasets and automate back-office work that previously required more manual effort.

“The way we’re using AI is on volume data,” Bevan says. “It’s great if you’ve got mass market products, big volumes. It’ll analyze that super fast. It’ll give you real time.”

But Bevan adds, “I actually prefer the actual customer feedback.”

THE LOUDEST VOICES CAN MISLEAD

Customer feedback has a basic problem: not everybody gives it. People who have had an unusually good or bad experience may be more willing to write a review. Many satisfied, indifferent, or quietly disappointed customers simply move on.

Even the way companies collect feedback can affect what they hear. Qualtrics XM Institute’s 2025 research, based on 23,730 consumers across 23 countries and regions, including the UAE, found that feedback shared through surveys tended to be more positive, while feedback given through phone calls and emails tended to be more negative. 

That creates another challenge for businesses: customer sentiment can look different depending on where they listen.

Trip.com tries to reduce that risk by looking at reviews alongside broader behavior, including searches, bookings, cancellations, repeat behavior, and customer service interactions. 

The company also looks for patterns across markets, customer groups and time periods rather than reacting to an isolated change in sentiment. 

“A highly vocal group may highlight an issue that is important, but we need to understand whether it represents a broader structural trend or a specific experience affecting a particular segment,” Sudhakar says. 

Bevan approaches complaints from another angle. He recalls advice from a former website manager that one complaint could point to many other customers who experienced the same problem but never took the time to report it. 

“If one complaint comes in, you should do something about it because that means there are 100 others who haven’t bothered to take the time to complain. So every complaint, if it’s negative, especially, we have to take into consideration because it means there’s a problem somewhere, or we’ve missed something,” Bevan says

WHAT PEOPLE SAY VS. WHAT THEY DO

Perhaps the biggest challenge to traditional market research is the gap between stated preference and actual behavior.

Sudhakar sees that gap in how travelers make decisions. Someone may say they want the lowest possible price, but eventually prioritize location, flexibility, convenience or a particular experience. “What is really valuable is being able to see the difference between what travelers say they value and what they actually do.”

That doesn’t necessarily mean the customer was wrong when answering a survey. Purchasing decisions involve trade-offs that can be difficult to reproduce in a research setting.

“Behavioral data tells us what people do, while surveys and qualitative research can help us understand why they do it,” Sudhakar says. 

If bookings show travelers choosing more short-haul destinations, the data reveals the change. It doesn’t explain whether customers are motivated by convenience, connectivity, value, or a desire to travel more often.

Those explanations can lead to very different business decisions. Ongoing feedback shows what’s changing, while traditional research helps explain why. More and more, the two work together instead of replacing each other.

HUMAN JUDGMENT STILL MATTERS

There is another limit to what customer data can capture. Bevan says human interaction remains particularly important in tailor-made and luxury travel. 

Customers may build relationships with agents who understand their families and preferences, then simply message them on WhatsApp when they want another holiday. 

That knowledge is personal. It develops through repeated conversations, not simply a record of clicks.

Sudhakar makes a similar distinction around AI. Trip.com uses the technology to help customers find information and support travel planning, but he says some situations still require human judgment, particularly when trips are disrupted or customers need reassurance. 

The companies that understand customers best may not be the ones collecting the most feedback. They will be the ones who know when to trust what customers say, when to watch what they do, and when to ask why.