Artificial intelligence is transforming global travel experiences: Agoda

The tourism industry is complex and globally distributed, requiring careful consideration from consumers who are presented with a huge amount of choice when planning their travel. Small decisions, like room preferences, can significantly impact the travel experience.  
 
At the same time, travel businesses and service providers struggle to reach a global audience and clarify their offerings.  
 
Digital travel platform, Agoda’s Chief Technology Officer, Idan Zalzberg in an exclusive interview with iTNews Asia, said digitalisation is unifying and standardising travel services, with innovative technologies like generative AI further simplifying processes to provide greater personalisation. 
 
“At Agoda, we have seen a lot of success in treating the attention of the consumer as a precious and limited resource and designing our software to bring forward what each consumer cares about,” he said. 
 
He added the company is leveraging generative AI to improve internal efficiency and customer experience, with 60 projects currently in use. 
 
The company’s approach to AI chatbots for consumer-facing applications is what Zalzberg calls ‘bottom-up’: focusing on doing one thing well, such as addressing specific customer questions about a current property, rather than trying to solve all problems.    
 
With the “connected trip” approach, Agoda aims to deliver a unified experience, enabling customers to book various aspects of travel in one place.   
 
The challenge was to merge and unify different product funnels, such as accommodations, flights, and activities, into a cohesive platform. It needed to generalise different interfaces to handle multiple travel products in a unified way.  
 
To achieve this, Zalzberg said they made significant changes in data management and technology adoption and worked on building a more flexible and scalable architecture.  
 
His team of 1,600 developers has followed a data-driven approach, achieving a 30-fold increase in data volume and adopting a rapid ‘test and learn approach’ to improving the product.  
 
Transitioning the tech stack from .NET to Scala and Kotlin, the firm has developed an experimentation platform to quickly test and validate changes across the entire platform.  
 
Zalzberg said the experimentation platform is across the board, with capacity to run about 1,000 experiments at any given time, meaning that no two people will ever see the same Agoda website, with the average customer exposed to 70 experiments at any one time.  
 
Improving developer productivity 
 
Agoda is leveraging observability and automation to improve developer productivity, with an ‘inside out’ approach to AI adoption which sees the majority of gen AI used for internal productivity gains.  
 
The company has built its own metrics and logs tools to improve observability across their systems. Zalzberg said he has a specified DevOps team that focuses on building deployment systems, service catalogue, and other platform engineering tools to enable developers to work independently without needing approvals. 
 
Kubernetes-based infrastructure has enabled Agoda developers to just describe what they need and have it automatically deployed across multiple data centres and clusters. This migration has decoupled developers from infrastructure teams, enabling faster deployment and maintenance, allowing Agoda to run tests at scale without sacrificing velocity. 
 
 “We have also built a chatbot that developers can use to get help with things like writing SQL queries, summarising threads, and getting information from the company's documentation,” he added.  
 
Agoda has adopted 2,000 microservices to work autonomously. “Ultimately, it’s designed to speed up our people and empower them to do more.”  
 
The metrics are collected and run through an in-house analytics program, enabling a holistic view of how each test impact the user behaviour. The growth of travel data allows us to plan, predict market trends and optimise our own marketing and operational work to keep the business lean and efficient, he added.