Since the mid-1950s, technology, business, and academic professionals have been intrigued by the concept of Artificial intelligence (AI). Fast-forward 60+ years and we are now seeing commercial implementations in various sectors, including insurance. So, will risk and insurance professionals embrace this development?
Theory and layers of reality
As with most worthwhile concepts, there is no simple description of AI. However, there are generalized constructs that can set the stage for analysis.
For example, in the session “The Use of AI in Insurance” at the 2017 Insurance-Canada Executive Forum (#ICXF2017), Jamie McDougall, Vice President of Business Intelligence and Analytics at Gore Mutual, provided a diagram defining AI by its components:
Does this mean that machines will take on functions currently handled by human knowledge workers? In short, yes, with big caveats.
The biggest challenge is that AI requires strategic thinking, focus, management, and willingness to fail.
Where to start?
This is up to the insurer’s priorities, depending on the impact it will have on their target clients. A recent post in Forbes by Blake Morgan – a speaker, author, and customer experience futurist – has some advice.
Morgan notes that most people are not keen on contacting insurance organizations because “customers typically come away from their interactions disappointed and dissatisfied.” To address this, Morgan contends that AI “has the potential to disrupt the entire industry and greatly improve the insurance customer experience.”
Morgan contends that AI can be applied to improve the process, arguing that AI technology can implement “touchless” claims, eliminating multiple hand-offs, delays, errors, and frustration by clients, intermediaries, and insurers.
Moreover, Morgan notes that AI can contribute to areas beyond claims: underwriting, marketing, and data. management Success satisfies customers for the present and provides opportunities for new products in future (Morgan cites Telematics as an example).
All good, but the problem is that many insurers are not accustomed to running parallel ‘bet the business’ projects. Coming to a decision on where to start first could be a multi-year project in itself.
InsurTechs step in to lead
AI typically requires large computing resources to address Big Data sets and timely responses to complex transactions. And, as the discussion above demonstrates, there are opportunities to start using AI in a number of business areas. So how to choose?
Perhaps InsurTechs can help focus this discussion.
Max Kraus, posting on the LOGIQ3 THiNK Blog advises that “The hottest category of Insurtech in 2017 has arguably been artificial intelligence (AI) as 5 of the largest 15 deals have gone to AI startups.”
Kraus cites Digital Footprint – a start up from the UK which recently partnered with MetLife to utilize AI to “access a previously vastly untapped data stream: social media.”
The objective – assuming client permission – is
to trawl through sites like Facebook and Instagram to flag information that can be used to determine risk factors, such as an applicant’s physical shape and smoking habits. In addition to identifying these risk factors, the data gathered by Digital Fineprint’s AI gives the insurer the ability to effectively manage their relationships and the potential to unlock countless other cross-selling opportunities.
This approach – starting small, preparing to fail fast, and leveraging success – controls risk. But, from a corporate perspective, this can’t be a serial project plan. Multiple initiatives are required to allow for some successes to balance failures. So, the enterprise requires careful management of multiple, parallel initiatives.
Do we dare?
That said, AI has significant upsides and insurers ignore it at their peril. A March 2016 PwC report – AI Insurance:Hype or Reality – makes this point:
AI’s initial impact primarily relates to improving efficiencies and automating existing customer-facing, underwriting and claims processes. Over time, its impact will be more profound; it will identify, assess, and underwrite emerging risks and identify new revenue sources.
The other panelists at the #ICXF2017 session – Bhanu Kholi and Jason Theodor from Capco – and the moderator – Catherine Kargas, Marcon – were sanguine about AI in insurance, assuming that insurers are prepared to manage Big Data and access to talent.
What do you think?
Editor’s note: We recognized that Artificial Intelligence may have significant impact on knowledge workers with job changes and possible displacement. We will address this in a forthcoming post.
