The key to the treasure chest
Data science at Allianz Asia Pacific
Perspective / /
Wan Ting Poh could have worked for Apple. The offer was on the table, all she had to do was to sign it. It is thanks to Raymond Au that the young data specialist ultimately decided in favor of Allianz. Au, a friendly gentleman in his early-50s has a gift for sharing his vision.
The data science team in Singapore
ou have to be pretty certain you're doing the right thing when you advise someone against working for one of the world's leading technology companies and entice them to an insurance company. Well, if he had doubts, Raymond Au probably wouldn't be the right man for his job. Au, who has a degree in IT, has been managing his Singapore-based team for a year now – a team like no other in the Allianz universe. He wants to unearth a treasure that has stayed undiscovered for a considerable period of time: millions of data records stored on the servers of Allianz’s Asian group companies, which could disclose a heap of information about its customers. Au is certain that his data scientists hold the key to this treasure chest.
Even though all information are anonymous and don’t allow any conclusion about a person’s identity – no names, no addresses, no telephone numbers, no emails are given – a peek into the databases can tell you which products can be offered to which customer group at which points in time, and what the success rate might be, says Au. “Without knowing the identity of a single customer we still can predict how likely it is that a defined segment of the population would buy a certain product.”
The computer can identify patterns and calculate how likely an event is to occur if a certain constellation is in place. For example, the machine brain can work out that a customer who purchased product A is more likely to add product B to his shopping cart than product C, especially if he is 32 years old, has a medium-sized income, lives in a new house and has two children. Instead of stumbling around in the dark like in the past, agents can now target policyholders who according to statistical data are inclined to purchase follow-up products. "We give agents the tools that can predict the success rate of a sales offer," explains Au. "Malaysia General is the first company of Allianz in Asia to put this tool into practice."
However, the range of possibilities for using the databases is a lot more extensive. Data can also provide an indication of which components tip the scales for a customer to recommend a company to his friends or family and which are of secondary importance; the combination of place of residence and place of work makes it possible to tariff a motor insurance policy that is adequate to the corresponding risk; evaluating Facebook, Twitter and LinkedIn posts could show which insurance agents out there would be a good fit for Allianz. And the list goes on. "There is a plethora of opportunities to make the data speak," says Au. "Up until now, we've barely scratched the surface."
Since starting their job a year ago, the eight data scientists have already launched 32 projects, both at Allianz Asia Pacific’s headquarter in Singapore and across nine group companies in Asia. 31 of those projects have been implemented – with a savings and revenue potential of around two million euros. Only one didn't make the cut. Au doesn't think this is too bad a result – and he's not alone. Chief Digital Officer Robin Loh, who has the task of forging ahead with digitization for Allianz companies in Asia has a high opinion of the Data Science team too. "They have a key part to play in expanding our business," says Loh who was previously in charge of setting up the digital financial platform for Chinese insurance giant Ping An.
Despite all that, not everyone welcomes Au and his diverse team – made up of Singaporeans, Chinese, Thais and even one Australian with Vietnamese roots – with open arms right from the start. Au avoids using the word "resistance" – he prefers to call it a "challenge". Still, the fact of the matter is that in various group companies, he has his work cut out trying to convince them. And sometimes he has to explain that having someone like him and his team on board, people without an insurance background, might actually be an advantage. "Insurers can learn a great deal from developments in other sectors of the economy," he argues occasionally. And one of those would be systematic data analysis.
Efforts are already being made to consolidate the data of all Allianz companies in Asia in one system. "That's the dream – to have this kind of central data warehouse where all information is gathered automatically," raves Raymond Au. As things stand at present, data still have to be collected from each national subsidiary individually.
Incidentally, a systematic approach makes it not just possible to identify patterns but deviations as well. Data scientists call it "anomaly", everyone else calls it "insurance fraud". A computer can hit the jackpot in seconds and find inconsistencies within a claims file – inconsistencies that a claims administrator might overlook. The machine doesn't care whether it has to trawl through health insurance data or road accident reports. Ever since Allianz Malaysia Life decided to send its health insurance files through the digital raster, it has the capability to avoid future cost of 300,000 euros a year.
An argument that should eventually convince the skeptics too. As far as Au is concerned, however, there is something more important than that. And that is trust. "We're not some outsiders trying to tell other people how to do their job," says the 52-year-old expert. "We like to think of ourselves as partners helping others to solve their problems." These days, one of his team members sits at the table at Malaysia Life every 14 days – every time the next claims workshop is on the agenda. Au would like to see other partners be this accepting.
Robin Loh has the task of forging ahead with digitization for Allianz companies in Asia
He says that data science is the easy part of his job. The tougher question for him is how a new methodology can be integrated into the working process and how to get people on board with the idea of transformation that will change traditional operations and even make some of them redundant. It's human nature to want to hold on to and preserve something that has stood the test of time. Au gets that. But progress is never sentimental. It moves forward relentlessly. This is why the analyst is adamant his job is more than just developing the foundations of a digital business model together with his team. He wants to be there when that model gets implemented – in the thick of things, not as an onlooker.
To do that, it takes stamina. But that's something Au is familiar with. Every evening, he comes home, sits down with his daughter and goes through her homework. And whenever the eleven-year-old doesn't quite get the mark she had been hoping for, even though she'd worked really hard, her dad has to make her feel better. "All I can do is teach her something that's always helped me in life," says Au: "Even if you work hard, there's no guarantee you'll be successful. But if you don't work hard, you will certainly fail."
"The job of a data scientist is to use large quantities of data to generate information and derive recommendations for action, both of which should make it possible for the company to improve its efficiency. In order to do this, data scientists use innovative analytics tools and develop queries that can filter out valuable information from overwhelming quantities of data. This is followed by making hypotheses, which are then statistically verified and set up as decision-making templates for the management." (Wikipedia)