March 13, 2019 | Written by: Leo Frey, Head of Digital Platforms, Allianz Taiwan
Share this submit:
In accordance with Investopedia, monetary know-how has grown explosively because the cell web and smartphone revolution. Initially referring to pc know-how utilized to the again workplace of banks or buying and selling corporations, “fintech” now describes quite a lot of monetary actions, from depositing a examine with a smartphone to managing investments, usually with out the help of an individual.
Headquartered in Taipei, Taiwan, Allianz AG is a life insurance coverage firm that gives personal and company clients an array of life, accident and medical insurance merchandise.
As a result of the inhabitants in Taiwan is so tech savvy and eager to attempt new issues, Allianz always strives to innovate to maintain up with the calls for of consumers.
We at Allianz knew that synthetic intelligence (AI) know-how can be a solution to improve the shopper expertise and thus created a cognitive digital assistant referred to as “Allie.”
A digital assistant any time
Allie is fluent in Mandarin and handles 80 % of the inbound buyer requests to name facilities for Allianz. Callers can ask nearly any query associated to insurance coverage insurance policies at any time of day. Allie is not only a database for FAQs, she will additionally deal with coverage adjustments. Her solutions keep away from insurance coverage jargon and complicated terminology. Allie explains the solutions in easy ways in which folks perceive whereas following rules and authorized compliance tips.
We’re focusing extra now on the human part. To try this, we’ve included some small speak capabilities. Many purchasers start their chats with questions like “How’s the climate?” and we predict that enabling Allie to have pure conversations with folks will drive social adoption.
Taiwanese folks admire Allie’s cute solutions to all these questions, however in the long run, it’s nonetheless an insurance coverage enterprise, so Allie retains the dialog skilled.
Working with IBM
We selected IBM not solely as a result of we imagine IBM Watson is without doubt one of the strongest AI engines available on the market, but additionally as a result of IBM developed a form of center layer, which permits our knowledge to stay on-premises.
When the shopper is engaged with Allie, the query that the shopper is asking goes to IBM Watson on the cloud. Watson understands the query, determines the intent, then transfers the request to the center layer, which both calls upon our back-end know-how or gives the reply to the shopper, however not by means of the cloud. Which means that coverage inquiry knowledge and coverage change gadgets are safer, which, from a authorized and regulatory perspective, is a really easy answer.
The center layer is open supply, so we are able to proceed coaching IBM Watson and including extra to the data base in addition to combine APIs all on our personal. IBM delivered intensive coaching and our IT and enterprise employees are capable of execute sure eventualities with out extra help.
So, for instance, if we have to make a fast replace, we are able to do that with none help from IBM. We aren’t completely dependent and as a substitute, we are able to work with IBM on extra complicated facets of the mission.
Our first discussions with IBM about creating a digital assistant had been round October 2017. We formally kicked-off the mission in November 2017 and launched on the finish of Might 2018. Our first minimal viable product really had a really broad scope with greater than 20 APIs built-in to get coverage inquiry and coverage change data.
We’re very happy with what we now have achieved with IBM in such a brief time frame.
Optimistic buyer suggestions
We launched with a big media event along with IBM final yr, and the suggestions right here was very constructive from each the market and clients. Allie’s buyer score is four.5 out of 5 stars, and the corporate’s Internet Promoter Rating (NPS) continues to extend, suggesting that Allianz clients would suggest Allianz to others.
One other key efficiency indicator (KPI) that we recognized, which was shocking, really, is that 45 % of our requests are available in earlier than and after enterprise hours. That is fairly a excessive quantity, and earlier than Allie, we weren’t capable of reply all these requests as a result of our name middle is just open from 9 AM to six PM.
The subsequent step can be to develop Allie throughout channels to supply digital assistant capabilities to brokers and bankers along with clients.