It is an exciting time to be in the insurance industry. This sector is on the verge of a technological transformation, with a major component being AI. The self-learning capability of AI systems allows insurers to create new product offerings across different geographies and customer segments and AI helps insurers to reduce costs, meet customer expectations, and stay ahead of competitors.
To help you get the best out of this technological leap, we would like to share the views and experience from some industry experts joining the panel “Ingredients of a successful AI – best practices as well as learnings from past challenges” from Sthlm Fintech Week 2021. Here we have gathered the best practices on how to do AI in insurance and what the different learnings are when embarking on AI.
Panel speakers:
- Mattias Fras, Group Head of AI Strategy & Acceleration at Nordea
-
Marcus Janback, CEO at Insurello
-
Liselott Johansson, CEO at Greater Than
AI can be fantastic, but it’s not for everyone and everything
Today, AI is a hot topic – the phrase is used freely across industries and “everyone is using it” – and businesses worldwide are interested in how to benefit from it. However, is AI really for everyone and everything?
We can all agree that AI is a fantastic technology, but for it to be so, you need to start with the business case and do a deep dive investigation to see where it could potentially fit into your business. It is a maturity step for the insurance industry, which is becoming more and more data– centric and data–driven. Utilizing the power of data could be done in different ways, from smart rule-based business logic to more advanced algorithms, to the end state of smart AI and machine learning. There is a step-by-step learning exercise in this that the industry needs to forego. AI can be fantastic, but it’s not for everyone – and not for everything.
Greater Than – a great example where AI works
Greater Than provides AI that calculates and price risk for the individual driver in real-time, helping auto insurance carriers to improve loss ratio and supporting automotive OEMs with behavioral–based pricing for new mobility. The AI future claims cost predictions, per the individual and aggregated for the entire sample. Their end users don’t have to wait for the actual results since their AI enables customers to pro-actively plan their finances.
Covid-19 forces businesses to shift their future perspective
The pandemic has accelerated the need for better real–time data and the need to be more forward thinking in your predictions is now increasing. Thus, new pressures make it hard for insurers to keep pace and adopt AI technologies because of long development timelines and traditionally high investment requirements. Many companies are now looking to innovate using new technologies to help optimize their AI investments or launch AI programs with limited resources.
Change is coming. How do businesses stay competitive?
There are many dimensions to this question. AI technology is evolving to become more standardized and accessible, and with that becoming more of a commodity. Thus, one competitive advantage will be to pick the right business opportunities and problems to address. Thus, it is much more about the people, the culture and the problem–solving skill, rather than it is about the technology since everyone has more or less the same technology. The question is how do you leverage that technology and what are the qualifications people must have to make the most out of it. Today, we are all competing on the rate of learning. But one thing is clear, if you want to succeed and stay competitive, you need to transform your business.
What are the practical challenges with AI – where do you start?
It is important to have a clear structure, a main goal and to set up different KPIs. You also need to collect the necessary and structured data.
Finally, you need a great team in place. Once you have a clear structure, a goal, KPIs, good and structured data and the right team in place, you can start to implement it.
Quality data
For AI to be successful, you need access to high–value quality data to get something out of the process. Today, we possess a lot of valuable information that could be used in the process. Key to this is a combination of standardized high–quality core applications that can feed you high–quality data, but also a more open mindset. To stay ahead of your competitors and be successful in this data–driven landscape, looking for partnerships outside of your ecosystem to access other kinds of data might be the way. You need an open mindset and you need to use the right data, rather than being restricted to using the data you already have.
The ethical aspects of AI
AI will have a significant impact on the development of humanity in the near future. It has raised fundamental questions about what we should do with these systems, what the systems themselves should do, what risks they involve, and how we can control these. Today, there is a debate about what constitutes “ethical AI” and which ethical requirements, technical standards and best practices needed for its realization.
A good thing is that AI really exposes the ethical aspects of how we do business – but we also need to manage that. In the future, there will be more focus on ethical standards and more AI platforms will have built–in ethical control frameworks. We see more guidelines and principles coming out driven by the need for algorithmic transparency and explainability, which is positive.
Looking at L&P, it’s a bit more sensitive, because optimizing the risk of driving is one thing, but optimizing the lifestyle is a totally different thing. There are a number of implications from the privacy and the integrity side when it comes to L&P.
The proactive way of utilizing information and giving proactive preventive recommendations is a bit more neutral than having all the risks embedded into the pricing model. So using preventive measures is the way to move forward.
