Advanced cognitive solutions are all about creating knowledge from data to expand the expertise of virtually everyone in the organization, continually learning and adapting to out-think the needs of the market.
Data analytics and artificial intelligence are set to transform the insurance industry, with a recent poll * finding more than 90% of senior executives believe it will drive change in insurance companies’ business models.
A mix of mind and machine is needed in this new approach, but the direction of travel is clear: insurance is the one industry considered to be the most exposed to the transformative power of machine algorithms (and the most behind the curve) compared to others like banking, government services, retail or energy.
Two-thirds (61%) of C-level business leaders in the research acknowledge their companies could rely on data analysis more and intuition less. They don’t consider their own organisations to be highly data-driven. This puts them at risk of being surpassed by their more data-driven competitors, given recent advances in technology and data and analytics techniques.
There’s a likelihood that more and more routine business decisions will soon be made by machines. But what about those that require creative judgment and collaboration? How do data and analytics inform these decisions? And how is decision-making changing?
The conclusion from the recent Insurance Analytics Summit held in London was that there are big changes converging right now in several areas at once: in terms of the insurance operating model, in terms of the external digital eco-system, with challenges for talent management and for leadership.
“This is the end of the golden age of the CEO,” commented Florence Tondu-Melique, COO Hiscox Europe. “The current problems in the industry are complex and they are technological, requiring a team of experts and the ‘leadership of the many’, a kind of collective intelligence.”
Technology meanwhile is fuelling new expectations of customers. Consumers are more empowered with a better understanding of what they are paying for. They are more self-directed so that traditional product pushing is no longer effective.
In terms of the digital eco-system customers want a seamless experience and they want the technology to be sufficiently powerful to be able to create warmth and emotion — in a world that is dematerializing and moving to digital experiences and digital satisfaction.
“In terms of the client relationship, all of this is creating customer churn at the moment,” said Florence Tondu-Melique. “We are seeing an explosion of data, which is our historical battlefield as insurers. The challenge is in taking structured data and turning it into insights…The future is in industrialized personalization.”
Looking ahead, in terms of the insurance operating model, there is a need to drive efficiency with greater automation (for example in claims, fraud detection and underwriting) and digitalization generally. A recent study ** predicts that insurance will see consolidation of 25% of its full-time positions, with many existing roles becoming highly data-driven. It estimated that 60% of a typical underwriter’s job can be automated, creating a premium on certain types of new data skills, whilst others will decline.
“I am a big believer in business models that combine both the strengths of humans and AI,” commented Monika Schulze, Global Head of Marketing, Zurich Insurance in the conference.
“I believe AI is the future for the insurance industry but we have to be clear what we want to achieve…There is a huge potential to deal with human bias, for example in some claims processes. AI is here and it won’t go away. For us at Zurich, one of the challenges has been getting people into the mind-set that the technology can be positive. In the beginning we had so many ideas and we had good business cases but after one or two years some of the numbers got lost. The big thing is to focus. Set up pilots, make it clear and stick to the original plan.”
There is a traditional insurance mind-set that a sample of less than a million of users is insignificant as a test. But insurers are getting much better at adopting an agile methodology: take 100 people and test before going on to 10 million users.
With super-computing and big data repositories like Hadoop and HPCC Systems now available as a vast distributed file system of consumer data and unstructured records, insurers are in a much better position to understand the deeper ‘context’ of customers’ decisions.
With the new platforms available, a cognitive business has systems that can enhance digital intelligence exponentially, supporting underwriters and the sales force.
Understanding the ‘context’ of language involves not only converting a call centre conversation into a text record in sub-second, but also scoring it against any number of risk scores that may be required, for example detecting the likelihood of deception in claims as defined by things like voice tone, over-talking, hesitation, repetition or use of pronouns.
Context, in terms of the intent of the customer, is what it is really about.
People sometimes lie of course, so there are questions for advanced solutions in connecting language to a personality profile or a risk profile as a set of personality traits.
Cognitive machines can already detect emotions in text or in a call centre log. Looking into the future this will involve looking more and more into the context and the intelligence behind each customer interaction. There is great potential for insurance in this area, being able to deliver rich data to insurance managers working in the claims, fraud or underwriting system, but also allowing them to look transparently into the context of how certain scores are arrived at, rather than just delivering them.
The challenges involve getting access to the many new external data sources, as well as the cognitive approach to processing the data itself.
Follow the link to the LexisNexis Risk Solutions website to find out more about how we support insurers.
The Insurance Analytics Europe Summit was held in London 5-6 October 2016.
** McKinsey’s ‘Unleashing the Value of Advanced Analytics in Insurance‘