SITA Quickly Finds and Returns Lost Items to Airline Passengers With Real-Time Tracking

Airline passengers lose millions of items across thousands of aircraft and hundreds of airports every year. These losses are emotional and distressing for passengers and expensive for the aviation industry. It can cost up to $95 to locate and repatriate a single lost item. Multiply that by millions, and the industry spends billions annually on making it right.

SITA exists in part to solve this problem. Founded in 1949, they are the leading global specialist in airport information technology. SITA facilitates inter-airport communications and tracks lost and found items on planes and in airports—a separate issue from lost baggage, which is usually tagged at check-in. 

There are many variables to consider when reuniting travelers with lost items. A passenger may not know where they lost the item or may describe it differently than how the found item was logged. There can also be hundreds of potential matches—consider the uniformity of mobile phones—and ensuring data privacy adds another layer of complexity. 

SITA faced a connected data challenge—they needed an easier way to match items reported lost with those reported found. SITA realized that the relationships between data points are just as meaningful as the data points themselves, which is why they chose Neo4j’s graph technology to power a new application that tracks lost items, according to Aidan Fries, a lead developer at SITA. 

Making the Right Connections With a Graph Database and the Lost and Found Property (LFP)

SITA initially acquired and deployed WorldTracer® Lost and Found Property, a purpose-built, turnkey application to connect travelers with lost items. It was designed for venues and stadiums, but in the last few years, SITA began working with developers to extend its use for air travel. The tool evolved into Lost and Found Property, or LFP—a new application built on a graph database and developed under the WorldTracer® banner—now in use at ten airports around the world.

“The primary motivation for deploying LFP was to make the process of getting lost items back faster and less painful,” explains Fries. It used to be that a traveler would experience the panic of losing something, then go to a service desk to ask for help, beginning a complicated chain of human inquiry to determine if the item was found, and where. 

Technology helps, but using the right technology is vital. “Matching such large numbers of lost items with the descriptions made by passengers is incredibly demanding computationally and wouldn’t be possible with a static, relational database,” he says. The same applied to making inbound calls into SITA more productive: passengers would be better served seeing how this is tracked on an app, versus having to call for status updates.

The LFP uses a suggestion engine to calculate a score based on item properties and returns potential matches to an admin, who can then confirm.

“The process of generating these suggestions is only possible with a graph database,” says Fries. “Such a system needs to analyze and draw conclusions based on input about location, item specifications like size, shape, color, and a host of other parameters. With a more traditional relational database, creating those links is going to take much more time because of the static nature of the database’s design.” 

A traditional relational database, where data points are housed in rows and columns, doesn’t account for the relationships between the data points—attributes that a reported lost item and a reported found item have in common—whereas graph databases are designed to infer connections and rapidly establish relationships and links between billions of data points.

“What we were trying to do was quite complex,” Fries explains. “A certain level of expertise was required, but open lines of continuous communication were just as important. The support we received from Neo4j was unparalleled, and it’s because of our close collaboration that we were able to get this off the ground.”

“Partners like Neo4j are rare,” he adds. “Many vendors are competent or offer excellent customer service but rarely as much as Neo4j, and in equal measure.”

Time Saved and Minds at Ease Through Automated Processes and Accurate Item-Matching

The LFP application is easy to use. It now only takes agents 30 seconds to register a found item and a minute to complete a full report. Passengers can access the system from any location 24/7 via a simple web portal to report, pay for, and organize repatriation from any device connected to the internet. It’s the LFP’s accuracy, and quick responses to passengers, however, that provides a differentiated customer experience. 

“It’s quite remarkable what we’re able to do now and what our users can do,” says Fries. “We’ve drastically improved visibility over the lost property process.” 

The system also automates a large portion of the repatriation process, reducing the time agents in call centers previously spent assisting.

Finding Additional Uses for Graph Technology

As SITA looks ahead, their goal is to reduce the number of technologies used across the organization. SITA’s architecture teams are exploring how Neo4j might underpin more of their applications.

“We’re a big organization that entails lots of departments, many of which operate autonomously from one another,” Fries explains. “Ultimately, we want to harmonize the data and databases that departments work from.”

Get in Touch

Curious about what insights you could unlock for your business with graph-powered solutions? Let’s talk – reach out, and we’ll get in touch.