Celine Le Cotonnec – Data Analytics in insurance

Celine arrived 16 years ago as a student in Sinology and is now Chief Data and Innovation Officer at AXA Insurance in Singapore. Read this amazing story!

Hi Celine, so what do you do and what brought you to Singapore?

I came to Asia 16 years ago as a sinology student in a Taiwanese university. Learning Chinese since I was 14 years old was the best choice I have ever made. After an International Business degree at Guangzhou University, I found a job at the French Consulate Trade Commission in Shanghai, supporting SMEs in heavy industries and new technologies to develop in China. This was my first experience in Tech, I will always remember that day in 2007, when I met a small Chinese company developing a third-party mobile and online payment platform. There were less than 100 employees at that time. It was nothing else than…Alipay!

With a business related background, starting by sinology, how come you ended up working in data and innovation?

In my previous role in PSA Peugeot Citroen China, I was leading the digital and media buying for our three brands: Peugeot, Citroen and DS:overseeing the performance of e-commerce strategy, analytics and social media campaign. Data Science was first used in Digital Media to improve retargeting, offer customized advertisement,increase traffic and improve conversion.

Later on, I took over the Innovation Department for PSA in China. We launched our first connected car and were looking at creating innovative services for improved driving experience based on car sensors’ data. The digital ecosystem in China is drastically different from Europe. Customers are younger and more connected. We had to develop a suite of services suitable for Chinese customers and find new business models with digital partners using car data to generate additional revenues. This was the only way to maintain expensive car connectivity and cyber-security infrastructures. It was also the beginning of Mobility-as-a-service. Everyone witnessed the emergence of platforms such as Uber in Europe, Grab in South East Asia or the Giant Didi in China. Similarly to what happened in the hospitality industry with the appearance of Airbnb, we saw business models switching from ownership to pay-per-use. Every car manufacturer was wondering how to ride on the wave of mobility services on top of the traditional car purchase offer. I was then asked to create a Business Unit on connected services and mobility to address those questions for the Chinese market.

So what does it mean to be a Chief Data Officer?

The role of a Chief Data officer is mainly to transform the organization to be more data-driven.Moving away from gut feelings to analytics enabled organizations to measure, track and monitor the performance of our processes, products, distributions and customer experience. Thanks to real-time visibility on our business, each employee of the organization can take fast and sound decisions, monitor the results, and apply corrective actions. Analyzing customer data helps understand their habits, market segment, life style, so it becomes possible to design customized insurance with the right coverage, sized risk, at the appropriate time.

I oversee a real diversity of data-related activities including data management, data quality and strategy for Singapore. The team is responsible for data analytics and business intelligence, which means creating valuable insights from data. We are also leading IT topics on data platforms related to business or customers. On the innovation side, our projects ranges from developing machine learning algorithms, implementing NLP or deep-learning technics to extract value from voice, images, pdf and web data.

What is your strategy in Singapore to transform an organization into becoming data-driven? What are the main things an organization needs to focus on when embarking on a data transformation journey?

Singapore in the leader in South East Asia when it comes to innovation and AI, thanks to the government –led Smart Nation initiative. To strengthen and accelerate data-driven transformation in any traditional organization, I would recommend to first focused on building four enabling pillars:

1. Platform and Tools:
Any data professional needs an environment to work: a Data Lake. Last year, we moved our various data infrastructure to public cloud in order to benefit from on-demand storage, computing and services. We have also moved to agile project management. While several teams were coding in different languages, the decision was taken to streamline every legacy analytical codes we had into Python programming language because of its simplicity, community support and numerous libraries. Finally, Tableau was widely deployed as visualization tool, speeding up decision making and KPI tracking. Anybody in the organization ranging from data scientist, analysts to actuaries, can now perform independent statistical analysis, advanced analytics, create and deploy machine learning models, at a minimum cost, with a competitive speed, and positively impact our customer experience.

2. People and Culture.
Changing an organization is not just about switching to new tools. It is first and foremost about changing the mindset of employees, their ways of working and raising awareness on what data-driven actually means. This year we set the ambitious target to train 20% of the organization in Python and Tableau. Data champions within a business unit, also called “Super Users” would  undergo an intensive four months training provided by the data team with a strong mentoring during the first months. Directors, and even Executive Committee Members would also undergo a 15 hours crash course in Python and data analytics. For the rest of the employees, there would be challenges on raising data awareness.

At a global level, we developed partnerships with e-learning platform such as LinkedIn learning and Coursera to encourage everyone to improve their  data and analytics capabilities.

Finally, we hold numerous events such as: lunch & learn with speakers from outside the insurance industry, panel discussions, evening meet-ups, sharing session with other data science teams from other companies. The main goal is to communicate on new business opportunities enabled by new tech and AI. We want to involve and empower the whole company to be part of this . Change Management efforts are key to achieve a true mindset switch.  

3. Data Governance
Using and storing data also imply compliance with data privacy and security regulations. Recent scandals such as Cambridge Analytica and the Singapore Data Breaches, remind us that large-scale data collection and usage could potentially raise significant privacy concerns. AXA is today the most forward thinking insurer globally when it comes to responsible AI and use of data. We are contributing to the public debate through collective actions such as IMPACT AI library In Europe or LiveWithAI think-tank in Singapore. A strong Governance framework is critical to balance between value and data privacy in the digital age.

4. Data management
It is nowadays a growing activities even in non-online businesses. Lot of people do not understand that 80% of a data science project is getting access, collecting, cleaning, and understanding the data. Predictive modelling that is supposed to be the most exciting part, is actually less than 20% of a data scientist daily job.
Value comes with quality and uniform data, as well as comprehensive guidelines for upstream users.
Once the basics are in place, it is a matter of weeks or even days to launch, test and industrialize a new AI. I’d like to quote here, Dr Deb Goswami, lead Data scientist at Traveloka, main online travel platform in Asia :” For a data science team, developing AI models is not the end game, but the value of the problem you are trying to solve”

According to you, how will AI disrupt the Insurance industry? And are the insurers afraid of Insurtech start-ups?

I wouldn’t consider AI as a disruption as it will only improve insurers’ efficiency as they become more customer-centric. The real disruption, in my opinion, will come from technologies such as Blockchain because it reduces middlemen such as agents/brokers, insurers, or even banks. In a trusted environment, people could pool risk among peers and get them directly reinsured without any.  

In the current value chain of insurance, it starts with the customer, of course, who buys an insurance from a broker or agent, the product is priced and underwritten by an insurer, who gets part of the risk reinsured by a reinsurer.

In today’s insurtech market, I would say that 80% of the start-ups are digitalizing the distribution experience, disrupting the intermediary but most likely supporting the digital sale of traditional insurance product.

The remaining 20% are working on solutions that would improve the efficiency of an insurer: AI in fraud detection, video-consultation to reduce healthcare cost and improve the customer experience, damage recognition from a car accident in order to speed up the process of surveyor and settle the claims faster with the counter-part insurer.

In the current context, insurtech start-ups are partners that can enable insurers to speed up their transformation and offer a better digital experience for our customers.

What do you find the most inspiring in the future you foresee?

The deployment of IoT and AI across all industries, made possible at affordable cost through the upcoming launch of 5G, cloud computing, and the emergence of blockchain will accelerate collaboration between platforms. While industries used to compete and work in silos, the new trend is to refocus on customer and collaborate to improved user experience through API integration. Technologies will also improve natural resources management and optimize existing assets.

Take the example of personal cars. In average they are  only in use 6% of their lifetime. Parking are expensive for both urban planning and users, and 30% of the traffic in big cities, such as Paris, are caused by people looking for a parking space. With the emergence of autonomous vehicle, the world will have to produce less cars and emit less greenhouse gas. In this period of fast change, every industry needs to transform and reinvent itself if it wishes to remain relevant in the connected AI-driven world of tomorrow.

How “dare you” be a women in data and how do you bring more diversity?

Diversity is not only about gender parity. Diversity is also about recruiting people from other industries, with different skill sets or culture. Diversity in ages and experience is also a great value inside and organization. Managing people with more skills or more experience should not be a threat.

I do support several initiatives promoting gender diversity in the data world. I am working with SheLovesData, girlswhocode and mentoring young female talents. In the book written by Sheryl Sandberg, Lean In, the issues of women in leadership are well described: putting others before themselves, lack of networking skills and being afraid of reaching out to their network or refusing a job because you’re not sure to have 100% of the skills required. I’m quite proud of the female ratio we have in our team or within the global data family. Our group lead data architect is actually a woman and there are several women CDO in the region I’m reporting to. Be bold, take risks, don’t be afraid and fight for your values are the advices I would always give to young mentees.  #daretobeafemaleintech

Celine, thank you ! 
Get in touch with us @ womenfrenchtech at gmail dot com
In collaboration with Amel Rigneau & Sophie Guo