Airport Eases Travel with Parloa AI Agent

Berlin-Brandenburg Airport is Germany’s third-largest airport, serving about 26 million passengers each year, with customer experience one of the major focuses of airport executives.

The airport’s call center could provide information about flights, delays, terminals, gates, and more, but it was very inefficient.

“We weren’t really happy with what it was delivering,” says Christian Draeger, senior vice president for passenger experience. “We didn’t always have consistent information being delivered. Some agents were more attuned to the latest information than others. We weren’t really happy with the cost, and we weren’t happy with the overall experience.”

Too often customers had long wait times, and they couldn’t reach anyone in the late evenings nor very early mornings, Draeger adds. “This is not the setup that we wanted to see. We wanted to have a much more technology-driven solution.”

With that in mind, executives started to look for solutions featuring generative artificial intelligence. Draeger explains: “There are a lot of flavors of AI out there. From our perspective, generative AI had the ability to provide the most intuitive conversations. We wanted to provide an experience that customers would pick up on.”

After receiving approval to pursue the project, the first step was to bring in a partner that understood the technology.

“Parloa is a leading company in this area in Germany and beyond,” Draeger says, adding that Berlin-Brandenburg executives weren’t concentrating on finding a German firm. “We were very much convinced by their approach.”

Parloa’s AI Agent platform provides users the control and flexibility to design enterprise-ready agents without getting boxed into static flows. The platform supports fully dynamic conversations, contextual memory, and access to any system, so agents can engage with customers rather than simply responding.

After choosing Parloa but before signing the contract, Berlin-Brandenburg brought in people from all parts of the organization to build the knowledge base that would serve as the backbone for the AI agent. “To do that, you need all different walks of life from the company,” Draeger explains.

Using the Parloa AI Agent platform was relatively easy; it took only six weeks to develop the AI agent. “There were a number of contributing factors that enabled us to move so fast. We had participation of all parts of the company in this project. We had people believing the project would add value. We also felt that the generative AI technology itself, after the initial training, would enhance the experience.”

The technology’s self-learning mode was also important because it enabled the airport’s AI agent, named the BER Agent, to continue to improve as it handled more passenger requests, Draeger adds. “If a customer called in one day and the AI agent didn’t give the perfect answer and then he called back two days later, through the self-training, the agent would have evolved and the answer would be a lot better. That gave us a lot of confidence.” Berlin-Brandenburg launched the BER Agent in the spring, and so far it has exceeded expectations. The technology handles initial calls, though some do still go to human agents.

Since installing the technology, customer satisfaction with airport services has increased to 85 percent from about 60 percent. Though pleased with that figure, Draeger is looking for a further increase. Call center costs have decreased 65 percent, even while handling calls at all times of the day and not just during business hours.

“This is constantly evolving,” Draeger says. “We initially handled 18,000 calls per month, and this continues to increase.”

The BER agent has also been invaluable during weather events and other crises because the technology can quickly be updated with the latest information.

In the year ahead, Draeger expects to add additional languages to the German, Polish, Spanish and English that are already available and to increase the personalization of the technology.

The Payoff

Since deploying the BER Agent developed on the Parloa AI Agent platform, Berlin-Brandenburg Airport has seen the following results:

• an 85 percent customer satisfaction rating, up from 60 percent;
• reduced customer wait times, from around two minutes to zero;
• passenger inquiries answered 24/7 in natural language; and
• a 65 percent decrease in call center costs.