Using AI based data insights in commercial insurance in the COVID-19 new normal

In the shadow of COVID-19 and its potential to change the way we live and do business, Dick Welch, President and CEO of Hospitality Insurance Group spoke with David Schapiro, Co-Founder at Planck, about the need for creating efficiency in the underwriting process and how AI based data insights can enable this.

Interviewer notes:

I had the honor of meeting Dick virtually in a COVID-19 roundtable discussion hosted by Planck. I had heard a lot about Dick and the innovative team at Hospitality Insurance Group, but this was my first opportunity to hear him discuss COVID-19 and its influence on the commercial insurance business, and on our lives. I was amazed by his insights and thoughts about our past, present and future. Dick graciously agreed to speak with us about creating efficiency in the insurance underwriting process. The following interview is an inspiring discussion about underwriting, data insights and the approaching new normal of our lives.


David Schapiro (DS): Could you please tell us about yourself, Hospitality Insurance Group (HIG) and your journey in the insurance industry?

Richard (Dick) Welch (DW): Hospitality Insurance Group is a small mutual commercial insurance company that specializes in writing bars, restaurants, social clubs and package stores. We write in seven states: Massachusetts, New Hampshire, Vermont, Rhode Island, Connecticut, Pennsylvania, and North Carolina. Our lead line of business is liquor liability, and we also write general liability and commercial property. As a specialist, our challenge is to know our niche and be better able to serve it than our competitors.

I am a veteran of the insurance business. I joined HIG in 2018 after a few years with The Concord Group and 25 years in the Travelers organization. Since joining HIG, I have been highly focused on improving the quality of our data and our analytical capabilities.

DS: Could you share with us what you saw as the key challenges in underwriting a small business, in the good old days before COVID-19?

DW: Although by definition small businesses are small, in the US there are many small businesses  – well over 30 million I believe – so there are many small businesses to insure, but each one will end up paying a relatively small amount for its insurance policy. On the other hand, a small business can potentially have a similar risk exposure as a large business – e.g. a small restaurant requires a similar cost and amount of underwriting effort to assess the risk as a large restaurant, albeit for a smaller policy. This means less income, so as a company insuring small businesses, we are challenged to achieve an underwriting profit.

Data on small businesses tends to be sparse, sometimes inaccurate, and lacking structure; while changes in small businesses are frequent and can be substantial. There are many different types of small businesses, and each type breaks down into sub-types. The accurate detailed business description is not always covered in SIC codes or other business classifiers. Trends can be harder to spot and aren’t always covered in current data points. 

As insurers, we often rely on self-reported data, but this tends to be highly prone to inaccuracy, inflation/deflation and potentially fraud.

DS: How has this changed with COVID-19, and how would you envision this in the new normal?

DW: Well, currently many small businesses are on hold, and it is difficult to assess how they will restart in the new normal. Some businesses, e.g. restaurants, bars, etc. will probably change – for example, with more delivery, table distance and outdoor seating – and this will impact the risk exposure and how we underwrite these businesses, requiring new data and insights that might not have been required in the pre-COVID-19 days. As small businesses are transformed, so too will the risks associated with them. 

It appears that some types of business models are being re-created on the fly. And as the new normal is a dynamic moving target, there will probably continue to be changes in businesses for the following months (years?), requiring dynamic data and underwriting capabilities.

DS: What role does data play in all this?

DW: There are a few key points, including the need to validate customer self-reported data, acquire conformed data that can be used in analytics and the need for additional data to underwrite small businesses; all of this while increasing the accuracy and coverage of our existing data points.

DS: How can we create efficiency in the underwriting process?

DW: The concept is simple: we can create efficiency by collecting accurate, conformed, actionable data that does not require underwriters to spend a large amount of time collecting. But transforming this concept into a process is a challenge, requiring advanced tools and data insights for underwriters and automation of many parts of the underwriting process. 

The accuracy and coverage of our current data sets need to be vastly increased, including their dynamic values. When a business changes, the data must change. We will need new data that currently does not exist or is not accessible, and new types of data will continuously become necessary.

The underwriter will need to focus on the actual assessment of the risk, and not on validating the data accuracy or performing a manual process. The data input and process will need to be automated and capable of keeping up with the dynamics of changing small businesses.

DS: Is there a way that artificial intelligence (AI) and other technologies could assist in creating this efficiency?

DW: AI and similar technologies are required to reconcile multiple sources of data to produce the most accurate insights possible and by exploring inter-relationships between underwriting variables. This is one of the primary components to create efficiency in the underwriting process.

DS: Could you please provide insight on how you are addressing this at HIG?

DW: HIG is engaged in a project to audit its entire liquor liability book of business on a risk-by-risk basis using Planck data. This study will reveal where underwriting data is potentially inaccurate and will introduce new variables to the underwriting process that will help to improve results.

DS: If you were to share one insight in these COVID-19 times with another insurance CEO, what would it be?

DW: Stay on top of what is changing in the business sectors you serve as a result of COVID-19, both directly and indirectly, and develop plans to adapt to those changes. While this may be a more difficult time to listen and communicate, it is probably more important than ever to do so.

Bio – Richard (Dick) E. Welch, Jr, President & CEO, Hospitality Insurance Group

A well-known and respected insurance professional in Massachusetts, Mr. Welch also brings a wealth of experience to his position – in particular, a strong and established track record of directing insurance operations, financial management and strategic planning.

Prior to joining HIG, he was most recently the ​Vice​ ​President​ ​of​ ​Corporate​ ​Planning​ ​for​ ​The Concord​ ​Group​ ​of​ ​​​Insurance​ ​Companies​ ​in​ ​Concord, New​ ​Hampshire. He is also​ ​the​ ​Founder​ ​and​ ​Principal​ ​of​ ​REW Insurance​ ​Consulting​ ​Services​ and was the ​President​ ​and​ ​CEO​ ​of the​ ​Premier​ ​Insurance​ ​Company​ ​of​ ​Massachusetts for 11 years.

In addition, he is also a ​former​ ​member​ ​and​ ​Chairman​ ​of the​ ​CAR​ ​Actuarial​ ​Committee,​​ Massachusetts​ ​Property Insurance​ ​Underwriting​ ​Association,​ ​and​ ​Center​ ​for Industrial​ ​Mathematics​ ​and​ ​Statistics​ ​at​ ​Worcester Polytechnic​ ​Institute.​ Mr. Welch has also served ​on​ ​the​ ​Board​ ​of Directors​ ​at​ ​Premier​ ​Insurance​ ​Co​ ​of​ ​Massachusetts,​ ​First​ ​Floridian​ ​Auto​ ​and​ ​Home​ ​Insurance​ ​Company, and​ ​First​ ​Trenton​ ​Indemnity​ ​Company.




Related Articles

We use cookies to provide the services and features offered on our website, and to improve our user experience. Learn more