Scaling customer service required Agoda to decide where AI could handle complexity and where human oversight remained necessary. To address concerns from customers and property owners more effectively, the company had to provide consistent assistance across communication channels.
“Scale and complexity are customer service’s biggest challenges,” Gauri Vicknasvaran, Senior Director, Customer Service Delivery, Agoda, told Frontier Enterprise.
Concerns often span multiple time zones, languages, and communication channels, and may involve several parties. Resolving them quickly and fairly requires effective tools and clearly defined processes.
At the same time, Agoda’s customers increasingly expect the company to understand their circumstances without requiring them to explain their situation repeatedly. Context, therefore, makes a significant difference to the customer service experience, Vicknasvaran said.
Pain points
According to Vicknasvaran, customer complaints commonly concern payments, booking changes, cancellations, and practical questions about properties. Accommodation partners’ issues, meanwhile, usually involve reservation changes and cancellation policies.
Human customer service agents work on shared platforms co-developed with Agoda’s product and engineering teams. This allows improvements to be rolled out globally instead of being introduced as isolated fixes, and supports collaboration with other internal teams.
“We have clear escalation paths across Legal, Security, and Data Privacy where needed, and shared dashboards give customer service, product, and marketing a common view,” Vicksnavaran explained.
AI strategy
For customers, Agoda uses a booking bot and a property AMA bot to provide immediate answers and direct users to the appropriate support processes. Instead of requiring every interaction to begin with a human agent, self-service tools can answer questions immediately and help customers navigate common issues.
The company also uses internal AI assistants to support human agents by reducing the time they spend searching for information before responding to customers. These assistants can summarise case histories, identify relevant policies, and help agents work through more complicated cases.
“AI is now a core part of how we scale support, but we always keep humans in charge of the outcome,” Vicknasvaran said.
This human-in-the-loop approach is particularly important when dealing with disputes, as an answer generated by AI may be technically correct without being sufficient to produce a fair resolution.
For irate customers and complex cases, Agoda retains human judgement to resolve issues within its policies.
“AI then helps behind the scenes by pulling together the history of a case and highlighting key facts and policies, so agents can respond faster and more consistently,” Vicknasvaran said.
During a tour of Agoda’s new technology hub in Bangkok earlier this year, CTO Idan Zalzberg discussed other uses of AI in customer service. Customer-service case information had previously existed only as text, limiting what Agoda could do with it.
“We have built a case summariser, because we found our human support agents spending a lot of their time just summarising what they have done. This was a waste of time that can be done automatically for them, and when they come into a case, the AI will give them the summary of what’s been done and tell them what they need to do next,” he recalled.
Agoda has also developed AI-based language translation capabilities, so human agents no longer have to wait for someone who speaks a particular language before responding to a customer query.
“By having real-time translation, we can substantially expand our ability to manage cases,” Zalzberg said.
Autocomplete and drafting tools also help human agents communicate with customers quickly and concisely by expanding their input into polite, clear messages.
“The customer will still get clear sentences that have empathy,” Zalzberg said.
Measuring the impact
Agoda tracks a broad range of customer service metrics, two of the most important of which are case-resolution time and customer satisfaction.
AI can affect both. Improved self-service may allow customers to resolve straightforward questions without waiting for an agent. Internal AI assistants can also help human agents process more complex cases by providing faster access to relevant information, Vicknasvaran said.
For Agoda, the value of AI in customer service is not simply the automation of conversations. It also lies in building infrastructure that can accommodate growing scale and complexity while helping employees make faster, better-informed decisions.
“Our aim is to surface issues early, set clearer expectations upfront, and resolve conflicts quickly and fairly,” she concluded.
