Key Takeaways

• Agentic AI is often more about talk than production services.
• Smart professionals focus on use cases and supporting tech.
• They test processes, refine the approach, and seek new opportunities.

Conversations with digital and business leaders about agentic AI often revolve around a similar sentiment: they have explored agents, but nothing is in production yet. However, while everyone talks about AI experimentation, no business can afford to run endless pilots without creating real value. With experts suggesting that professionals who fail to exploit AI risk being left behind, there is an imperative to deploy successful agents sooner rather than later.

At online travel specialist Booking.com, Huy Dao, director of data and machine learning platform, is tasked with delivering value from AI, including agentic services. He has produced tangible results by taking a structured approach to service rollout, creating targeted solutions to the challenges customers face today and tomorrow. Dao refers to this approach as the "connected trip," in which Booking.com attempts to ensure all elements of a customer's journey—whether flights, hotels, or attractions—are considered as an integrated experience.

Creating the connected trip means working across disparate information. The data stack Dao's team has built has allowed Booking.com to develop new AI-enabled services, including the company's first agentic application: a partner-to-guest system that facilitates communication between customers and hotel partners. This system, known as Smart Messenger, uses agentic AI to help hotel staff respond to guest inquiries more quickly and accurately. Booking.com reported that early experiments yielded a 73% increase in partner satisfaction compared to previous messaging tools.

1. Identify a Business Challenge

Dao emphasizes that the key to exploiting emerging technology is finding the right use case. While some professionals remain unsure about the potential of AI, he argues that companies can leverage agentic technologies to overcome intractable challenges. "AI is not like a flavor of the day, or even the year—it is the real thing," Dao says. "I see that every day at work, how AI can impact the way we do things."

At Booking.com, Dao and his team identified that timely responses to customer inquiries were a major pain point for hotel partners. They recognized that agentic technology could help hotels reply to questions faster and more accurately. Before the rollout, when a customer wanted to connect to a hotel partner—for example, to check if a property had a pool or to request a late arrival—the customer would contact the partner directly. However, hotel staff often needed to perform additional research to craft the right response, and they might be unavailable when the customer asked a question. This meant it could take several hours or more before the customer received an answer.

2. Build a Data Platform

Dao explains that the data stack his team has created allows Booking.com to accelerate the adoption of AI and machine learning technologies for use cases like the one outlined above. The Snowflake data platform forms part of an integrated stack that includes ThoughtSpot for analytics, Astronomer and Airflow for orchestration, Immuta for access control, Arize for machine learning observability, and AWS for cloud computing. The data team also tests and uses AI models from major providers, such as OpenAI, Amazon Bedrock, and Google Gemini.

Booking.com's bespoke partner-to-guest communication system was developed internally in Python, and the data team used LangGraph, an open-source agentic framework, to help the agent reason about guest inquiries. Dao stresses that effective agentic systems are not just about backend infrastructure; his team also thought carefully about the user interface. "We want to integrate technologies or AI capabilities wherever it makes sense to our users," he says. "In this use case, our partners already had a web-based portal to view their messages, so it was clear we should integrate the agent right there to help them."

This platform-centric approach is critical. By building a robust data foundation, Booking.com can now quickly experiment with new AI models and deploy agentic solutions without reinventing the wheel each time. The modular architecture allows the company to swap out components as the technology evolves, ensuring long-term scalability and maintainability.

3. Test the Use Case Carefully

With a business challenge identified and the technology platform in place, Dao and his team focused on implementation, which occurred in two phases. In the first phase, they developed a trusted assistant to help hotel partners deal with customer questions. The result was Smart Messenger, an agentic tool that gathers partner, property, and reservation information to support hotel staff in communicating with guests. In this initial phase, the human is still very much in the loop. "We want to make sure the partner is the one who has the final say on how they want to respond to customers," Dao explains. "But we give them an assistant, so that instead of taking five minutes to respond, it might be just a one-second click if they are happy with what the agent provides as an answer."

Dao emphasizes the importance of rigorous testing. His team conducted extensive A/B testing to compare agent-assisted responses with manual ones. They measured metrics like response time, accuracy, and partner satisfaction. The results showed that even with human oversight, Smart Messenger drastically reduced response times and improved consistency. The key was to design the agent to be helpful without being intrusive, allowing partners to maintain control while benefiting from automation.

4. Delegate as Confidence Rises

Over time, Dao says, confident hotel partners can start delegating more work to the agent—this represents the second phase of the implementation. Here, Booking.com's Auto-Reply tool allows hotel partners to define custom replies and create instant responses to guest questions, such as whether a hotel has on-site parking. "This phase is where the agent says, 'OK, if you trust me enough, I can act for you,'" says Dao. "In this use case, the partner might be sleeping when the customer asks a question, because it's late at night. However, the agent can respond on behalf of the partner—and that approach helps in several ways."

Booking.com reported that early experiments yielded the 73% increase in partner satisfaction. Dao notes that the agent continuously learns from past interactions and user feedback, adapting its responses for accuracy and relevance. "Now, with the agent, we measure the answer against everything we do; we experiment with it, and then we compare the improvement in satisfaction," he says. "Because the customer gets the answers they need, they don't have to contact customer support, and that success also reduces support costs."

The transition from assisted to autonomous mode is gradual. Booking.com provides partners with dashboards showing agent performance metrics, such as how many queries were handled automatically versus escalated. This transparency builds trust, allowing partners to confidently hand over more tasks over time. The company also offers training sessions and documentation to help partners configure Auto-Reply to match their specific preferences and policies.

5. Look for More Opportunities

Dao says agentic exploitation must be tied to the individual use case. As his team refines the customer experience, they continue to hone the platform, creating a foundation to support other agentic explorations. "We didn't want to build the platform for the platform's sake," he says. "When we built the platform, we had the user in mind. We made sure that we picked the right agentic technology."

Dao advises other professionals to take heed of the lessons learned. "When you do your testing, you might think the agentic system is good," he says. "But when you go into production, things like latency can become a problem that you need to deal with. Then, you must simplify your architecture and platform."

Looking ahead, Dao expects further pioneering developments at Booking.com over the next 24 months. "You should expect that, as a company, we will invest heavily in generative and agentic AI, not for the fun of it, but to increase the user experience," he says. "People are looking for a ChatGPT-like experience now, and we want to have a similar experience, or even better, when it comes to the travel experience on our sites."

The success at Booking.com offers a replicable blueprint for organizations eager to move beyond AI experimentation. By focusing on real business problems, building a solid data foundation, testing incrementally, scaling delegation gradually, and continuously exploring new opportunities, companies can turn the promise of agentic AI into tangible results. The 73% satisfaction boost is just the beginning—as agentic technology matures, the potential for transformative impact across industries becomes ever more achievable.

Source: ZDNET News