Ron Vermeulen is COO at Brightmaven, a company that specializes in services related to tracking down stolen and missing goods. By employing technology they make the investigations faster and more efficient. He was a speaker at the FRAUDtalks conference.
I would like to take you on the journey of a stolen object. A few years ago we were asked to track down a children’s merry-go-round valued at €125,000 that was stolen in Vienna. All the details of the merry-go-round were recorded, alerts were sent out to the Border Patrol and appeals were made on a number of platforms.
After extensive investigation we eventually found the merry-go-round in Sochi, where it was being used at the Olympic Games. We then notified the police in Moscow. They seized the merry-go-round, made investigations and caught the perpetrator. We went to the scene to collect the merry-go-round ourselves and return it to the Austrian owner. You can imagine how happy the owner was. This project took a total of 220 hours of investigation.
Every day insurance companies, lease companies, rental companies and individuals face damage caused by theft and fraud. Worldwide the damage amounts to 80 billion euro each year. That is a huge amount. Objects disappear from one place and resurface somewhere else in the world. Often with a different identity or in parts. And the cycle starts over and over again.
Stolen objects are moving in all directions: cars from Western Europe disappear to the Eastern bloc, but we also see cars that were stolen in Russia turn up in the Netherlands. The police force is overstretched and does not have the time and capacity for extensive investigations. Meanwhile the government is withdrawing and expects citizens and the market to take more initiative themselves.
Time for automation
That is why we have developed the platform Sjerlok: a combination of big data, Artificial Intelligence (AI) and old-school investigation. There is in particular a lot of interest in AI. It may sound very modern, but the first algorithms were already being used 50 years ago. Interestingly, the world is changing and digitizing: there is a lot of data and the entire world is within reach.
It would take decades to analyze all that data and filter relevant information by hand. Machine learning can help with this and speed up the process. By working in this manner we are faster, more constant and we are professionalizing the search for the stolen object.
How does it actually work?
Take the example of a stolen Porsche Macan. We make a detailed profile with information divided into three categories:
- Basic information (brand, model, color, license number, identification numbers of parts)
- Variable information (exact features, type of engine, transmission, interior and exterior)
- Unique information (minor damage, sticker, vignette, accessories).
Altogether there are around 80-120 characteristics that, together with photos, complete the profile. This profile is translated into 20 languages for different data sources. Then a 24/7 automatic search takes place on the internet and in databases such as online auction sites and forums. That way we can follow the criminal journey.
We also look on social media, not for people, but just for objects based on the profiles. And we collaborate with public and private organizations – including parties such as FRISS. In addition, there are interfaces that allow us to transport data in a multi-functional way, e.g. of car parks and border crossings, to see whether the stolen object has been there.
The ‘deep matching’ in the process means that valuable data can be extracted from all the information. In the future it can even be used to make predictions. This way of searching is more efficient, more constant, more cost saving and thus the losses are reduced significantly.
Lower costs, higher profit
Let us go back to the example of the children’s merry-go-round: it took 220 hours of investigation by hand. If we had had Sjerlok at our disposal back then, it would most likely only have taken 10 hours. All of a sudden it has become a lot more interesting for the client to retrieve stolen goods, as it is a lot cheaper. Based on our experience with Sjerlok and a shared services center we can guarantee a reduction in costs of around 15% of the total value of the stolen goods.
Some more examples: a boat costing 55 thousand euro that was stolen in The Netherlands, was found on a Polish auction site, seized and returned to its lawful owner. A Ford Pickup from the south of The Netherlands, also valued at 55 thousand euro, turned up on an auction site in Romania. Subsequently it was confiscated by the police in Bucharest, thanks to us.
It all sounds simple, but it is not. Because besides signaling where an object is, one also needs to actually find it, seize it, investigate the case and retrieve it. This requires a good international network to activate all the necessary parties.
A good example is a Bentley valued at 260 thousand euro. It was stolen in Amsterdam and after a long journey it was eventually found in Phnom Penh, Cambodia. That was mid-July 2017. After that it took around three weeks to convince the local police to seize the vehicle. The international network and the collaboration between public and private organizations are incredibly important for a successful result: the return of a stolen object to its lawful owner.
Every day we are busy tracing the journey of a stolen object, literally and figuratively all around the globe. To find stolen goods and return them to their lawful owner. And our primary motivation? Crime should never pay off. Never.