AXA Hong Kong is making artificial intelligence (AI) as a key part of its five-year strategy, using the technology on two fronts: automating specific tasks and as a bot used to help employees augment their intelligence in performing technical jobs.
“Within this timeframe, we expect all functions in the organization to have at least one AI engine, whether you are in finance, marketing, product development and distribution to name a few,” Xavier Lestrade, managing director, Life Insurance – Hong Kong & Macau at AXA China Region told Computerworld Hong Kong in an exclusive interview recently.
This will take the company’s commitment to deliver personalized customer service to the next level.
Starting the AI journey
According to Lestrade, AXA Hong Kong begun exploring AI from a technology perspective after some of his people attended the Dreamforce event organized by Salesforce in 2017.
In the first half of the following year, the company’s data team from the marketing department started building its first AI engine called “Next Big Offer”, which recommends the best protection for customers by matching the similarity between the purchase behavior of each individual and appropriate product features.
By taking the financial consultants’ feedback into account, the A.I. Machine Learning technique can fully perform smart learning to improve the forecast accuracy.
“We are using AI now to help us to calculate the next best offer, predict what is the next best offer and also predict when is the best time [to sell the offer],” Lestrade said. “What is important is to provide leads to the right customer at the right time.”
To develop this AI engine, AXA Hong Kong relied on its long-term technology partner Salesforce, which provided the insurance company’s CRM platform. The company implemented Salesforce’s Sales Cloud in 2014, the Services Cloud in 2016 and the Marketing Cloud in 2018.
“When I came to Hong Kong in 2012 from our Paris headquarters, I have this view that to transform a product-centric organization into a customer-centric organization, we need a CRM platform that enables us to have a 360-degree-view of our customers across all business lines and across all touchpoints from distributors, call center, social media and website,” Lestrade recalled.
Lestrade said: “So for a customer, when you talk to an agent in the morning and contact our call center in the afternoon, you can have a seamless conversation. You don’t have to repeat the information because data is shared across different touchpoints – thanks to our unique CRM platform.”
So when AXA Hong Kong decided to tap the power of AI, Salesforce entered the picture.
“Salesforce has an ecosystem that is enhanced every year with new solutions,” Lestrade said. “So this year, we implemented a powerful solution that leverages Mulesoft’s integration engine that enables us to get the data from all our backend systems and feed them to the Salesforce AI engine. And so when these data from legacy systems are combined with our CRM data that reside in the cloud, we are able deliver optimal insurance offerings to our customers.”
AXA Hong Kong rolled out its Next Big Offer, its first AI engine, in distribution and sales in July 2018 to more than 5,000 financial consultants.
“Before the roll out, we trained the financial consultants. The training was focused on user scenarios and how the AI-based solution helps them to be more productive and to understand their customer better,” said Kelvin Chung, head of Life Projects & Technology, AXA China Region.
Furthermore, a smart sales management platform called Lead Club was developed to offer an optimized matching between customers and financial consultants by using big data analytics. It is envisioned to increase sales opportunities and to ensure customers’ needs are best served by the most relevant financial consultant.
According to Chung, the AI engine gets smarter as the machine learns over time.
“We have a feedback system in our Next Big Offer because we have to keep training the model. Furthermore, our data scientists log into the model to fine-tune and more elements to the AI,” he said.
Evolving thoughts about AI
Lestrade said that AXA Hong Kong’s foray into AI is still at a very early stage and that they are “still learning by doing it”.
Attending the 2018 Dreamforce while the company was already in the midst of developing its first AI engine, he had a realization that change the way he view the technology.
“I was looking at AI more as a way to automate – as the next stage of automation,” Lestrade said. “What I realized is that insurance is a business of people and data. That is our fundamental business. We sell policies which are data. And we have quite a number of technical jobs that require high IQ.”
“I realized that with AI, we can augment the intelligence of our people who have technical jobs, helping them make the right decision in complex cases,” he said.
Using the technology to augment intelligence of people handling technical jobs is next in AXA Hong Kong’s AI pipeline. The company is hoping to extend this capability to its team of 20 underwriters by providing them with a personal digital assistant to help them make a decision in complex cases.
“What we anticipate is that there will be a very specific way to manage change and drive adoption by our underwriters of personal digital assistant, because we do not want them to have the wrong perception that the bot is going to learn their knowledge in order to replace them. Here, we are not looking for automation in these complex cases but we want to make sure we make smarter decisions,” Lestrade said.
The company expects to pilot the personal digital assistant to underwriters in the second half of this year.
Looking at AI beyond the technology
After last year’s Dreamforce, Lestrade came back convinced that he had to really look at AI as a key component of AXA Hong Kong’s five-year strategy.
“One of key messages that really spoke to me was that AI is a new resource and you need to manage the use of this resource as you have been managing human resources,” he recalled.
“And you will have some underperforming AIs, or eventually, AIs which might put your reputation at risk. So, how do you govern AI? You need to reflect on how to manage the performance of AI. It is not just implementing it and looking at it,” he said.
According to Lestrade, automating AI is not an issue anymore. “In most of the models we are using now, the machines are learning on their own. The question is how do you control that AI is not derailing?”
He pointed out: “AI, based on the data we provide, could recommend not to underwrite people due to racial consideration. Is this something we want to consider or not? But AI is only looking at data and the data might show there are different types of risk behaviors depending on people consideration. Is this something we are ready to accept? There might be some recommendation that we do not want AI to recommend because we cannot accept them from an ethical standpoint. This is a fundamental fact. It goes beyond technology, but it is extremely important.”
Bringing AI to the whole organization
Now, six months after rolling out Next Big Offer, AXA Hong Kong is seeing some benefits.
“We started with lead management because this is where we usually get the fastest return on investment,” Lestrade said.
To date, more than two-thirds of its financial consultants are using the
Lead Club, which has significantly improved their closing rate. Their rate of sale from AI-generated leads is higher than their self-generated leads.
Meanwhile, the company is starting to introduce AI into claims management.
“We are in the very early stage in our journey,” Lestrade said. “I am not sure we are across the 10% mark of what we can do with AI …. We are still learning and we want to make sure we have the controls in place to manage the performance of AI.”