Insuranceciooutlook

Future of Automation in Insurance

By Matthew Marino, AVP, Unum

Matthew Marino, AVP, Unum

In the time that has passed from when I was asked to share my views on automation I’ve had a surprising shift in my thinking. In my day job, I am responsible for our operational efficiency program which includes all things automation (RPA currently in broad use, NLP, ML, and AI in development for pilots and capability building). In the evening, I like many of you, study to stay on top of emerging trends most recently as a student taking part in MIT’s program “Artificial Intelligence: Implications for business strategy program”.

"The advantages of new tools in this field are the ease of development, the strength of capability"

What had I been thinking and where am I now at? Originally, I was impressed by the power and promise of these new capabilities. Even pessimistic views, while fearing the end of mankind as we know it, shared a view that massive transformation was in the very near future. I was more and more bought into how big of a change and more importantly how fast this change would become reality powered by better machines that could be taught to do so many things. Sharing a view that we need to be mindful of how roles evolve to best utilize the human skills and how to redeploy talent to these areas.

What has changed for me is not the potential of these tools so much, but rather my view on how rapidly we will be willing to adapt to leverage these tools to the extent they are capable. Put simply, if we look at recent comments made by visionaries such as Elon Musk and Stephen Hawking, AI may be one of the scariest new technologies to be introduced in recent history and when combining that with an industry that is by its very nature often risk-averse, one can take pause and see we as an industry probably aren’t yet ready to go all in.

An additional challenge I’ve observed in discussions with others across many different companies is how we look to these tools in the context of a strategic IT roadmap. One person I spoke with recently very honestly shared if we use RPA for short term gains, how will their strategic business case be negatively impacted as they look to consolidate and or replace their legacy systems? It was a very fair point, if we use the new tools on aging platforms as a band-aid to solve for the current challenges posed by having redundant legacy systems, we will continue to face ever-growing challenges on building the business case for integration/migration, is this a tradeoff we see our partners looking to make?

Given some of these headwinds, if I was to look out over the next 3-5 years what do I think I will see? I think some of these tools are easier to embrace than others, and I suspect many in our industry will look to these areas of comfort as a starting point on their journey. In my opinion, the use of robotics will become more and more widespread. Without taking away from the progress that has been made with these tools in recent years, the concept of desktop macros and even BPM in general, make a strong argument that applied RPA solutions is not a new concept but rather an evolution of practices that have long been in place. The advantages of these new tools are the ease of development, the strength of capability, but the general concept is not completely foreign and the idea that these can start off by being leveraged out of our IT areas as another developmental tool makes for an easier adoption I believe. 

NLP feels like the next level up in maturity. Again, building from an area of comfort, many organizations in the insurance vertical have long leveraged OCR tools and even in the last five or so years, speech analytic tools are becoming more and more widespread. The adoption I think we’ll see taking place first with more broad use of NLP capabilities will be that of a “digital assistant.” What I think we’ll have happening here is leveraging NLP to transcribe data into or out of our systems with human intervention for quality control. This notion of “control” is one factor that I think will cause the next tools that I’d look to in the maturity stages to be of growing concern.

For maximum value, when we look towards the use of machine learning, predictive analytics, and artificial intelligence I think we’ll see small initial use cases that organizations will start with, that will be so incredibly tightly controlled that much of the value is potentially diminished. I think this specific set of tools offers the most value and believe strongly that the organizations that structure the right balance between early adoption and pragmatic application will be truly those that differentiate themselves.

What conditions do I think will need to be in place for this to happen? I think the areas where we may see the best application for these tools happen to be the most sensitive. Sales, Underwriting, and Claims all come to mind for me as where the technologies can make the most value for an organization. I also recognize, that in the organizations I’ve worked these three areas are often the areas where words like relationships, judgment, and critical thinking are raised when discussing any type of process change. 

So how do I think the successful companies will overcome this resistance, I believe it must take place being led from a business perspective and not a technology play. Find a curious salesperson, agree to a small test, demonstrate the effectiveness (and the ultimate impact to commissions!) and get them to demand this. When we can prove value and be trusted partners, we’ll have a chance to succeed. If we lead into a group of sales folks with fancy language around deep learning, algorithms, propensity scores, and regression values we’ll likely get kicked out of the room. I think if we can learn from these partners a great deal about how to create demand, as opposed to how we in process excellence/IT can sometimes get so excited we forget to do this and instead try to force our way in. Always with the best of intentions but blind to the fact that until we can build demand we may simply be unwanted guests. 

When all is said and done, I think the company that will be most successful fastest will not necessarily have the best technology or the broadest applications, but simply the company that is best capable of managing change and leading the people they rely upon to a position where these tools are embraced and not rejected.

Weekly Brief

Top 10 Insurance Analytics Tech Solution Companies - 2019
Top 10 Insurance Analytics Consulting/Services Companies - 2019

Read Also

Match Game 2.0

Match Game 2.0

Clint Roszelle, Director – Enterprise Process Excellence, Citizens Property Insurance Corporation
Capitalizing on Commercial Insurtech Disruption

Capitalizing on Commercial Insurtech Disruption

Marcus Knuth, Vice President – Enterprise Technology, Acuity Insurance
Making Wiser Choices Regarding Technological Implementations

Making Wiser Choices Regarding Technological Implementations

Jill Cook, Director, Structured Settlements Administration, Protective Life
Driving Value-Based Experiences In Insurance  Using Apis

Driving Value-Based Experiences In Insurance Using Apis

Leslie (Les) Hermitt Jr. Chief Digital Officer C&F Digital Partners business unit
Impact of Operational Risk on Asset Valuations and Pricing of Corporate Insurance Policies

Impact of Operational Risk on Asset Valuations and Pricing of Corporate Insurance Policies

Aleksandar Kovacevic, Founder and Managing Director, Audeamus Risk
InsurTech's hot streak continues

InsurTech's hot streak continues

Sam Evans, Founding Partner, Eos Venture Partners