Insurance companies have always collected lots of data — now they are learning how to use it better. Vitor Ribeiro of AM RE Syndicate describes how the firm overhauled its own systems and processes to overcome its data challenges in-house.
What were the issues and how did AM RE approach the situation?
What is your advice to smaller companies in overcoming their data challenges and legacy issues?
What are some common sense and easy things companies can do to improve their data approaches?
What does the future hold for AM RE relating to getting the most out of data?
Data at AM RE
Regardless of size, insurance companies have always collected a wealth of data. Whether it’s a past policy holding firm many decades later (as long as the cause of loss occurred during the policy period), or an increasing number of policyholders moving online to review products and prices, make purchases, and file claims—the volume of data in insurance has increased exponentially in recent years. Case in point: over 90 percent of the data in the world has been generated in the last two years alone.
At AM RE Syndicate, a specialty program reinsurance provider, managing this increasingly vast amount of data was beginning to weigh heavy on resources. The company was faced with two options: to invest in external technology or to invest in its people. It chose the latter.
AM RE operates in a highly focused niche of the reinsurance market. It writes specialty quota share programs on behalf of reinsurers worldwide that want to participate in the US market. Since AM RE has agreements with cedants, brokers, sponsors, and reinsurers globally, its primary goal is to translate all the different data formats into comprehensive databases that can be used internally for analysis and externally for periodic reporting.
Re/insurers and, especially, smaller operations will often shy away from investing heavily in new technology and talent which end up enforcing a modelling or reporting “ritual”—a lot of time spent repeating formulas, pulling spreadsheets from folders hidden in the bowels of the filing system, formatting cells, and changing titles.
Off-the-shelf products often fit awkwardly, miss on vital components of the operation, cost a fortune, and do not provide the level of maintenance expected from something that can cost millions of dollars.
Faced with the grim prospect of falling into the same problematic pattern as other companies, AM RE decided to invest in its people and internally develop the expertise to tackle these common data challenges. Company executives made this decision when creating the foundation for AM RE due to its not having a single focus and having a team that is agile and small enough to be able to design solutions.
This proved to be most efficient for the company. Vitor Ribeiro, head of data at AM RE Syndicate, discussed the challenges and solutions AM RE has come across in its journey towards innovation.
What were the issues and how did AM RE approach the situation?
Insurance is an established industry riddled with legacy systems and in dire need of innovation. Although there is a myriad of tools and systems available to manage insurance data, if they are too rigid or difficult to integrate, they will never be successful.
As in any other operation, employees rarely have the same level of technical expertise—but if they are comfortable with Microsoft Excel, business intelligence (BI) software, or programming, the culture of communication and the approach on certain tasks can improve massively.
With that in mind, we developed all our automations to rely on user input and to accommodate each user’s “computer” skills. AM RE employees feel empowered when they understand what is being presented and how to work with the data. When it comes to external interactions, this capability is useful because it creates a dialogue with stakeholders where they are comfortable and competent.
Maintaining the integrity of information that comes in a variety of formats can be difficult but making documentation easily accessible is key. The Analytics Team is responsible for data integrity at the time of entry, and the values and documentation provided by the client are processed and inputted together, which provides the analysts with easy access to the audit trail.
Cross-team data integrity is enforced through further actions such as financial settlements and underwriting analysis.
What is your advice to smaller companies in overcoming their data challenges and legacy issues?
Encourage solutions that are unburdened by legacy thinking. Don’t overdesign processes and workflows. Understand the larger issue you are trying to solve and design a solution for that.
Most companies we deal with could form a pilot team to drive innovation and design solutions that increase productivity. From our experience with insurance companies globally, a team focused on operational improvement can have a tremendous impact on the efficiency of processes and quality of outputs. It is OK to fail if you can iterate quickly, so keep designs simple until you are on the right track.
Give the innovation team access to different levels of the organization chart because it is important to understand how problems translate from junior to senior employees and how decisions are ultimately made based on the information flow.
We have learned that being able to support each role within AM RE and keeping employees informed on all aspects of the company is key to being able to maintain a prosperous business. Implementing these same ideas is essential in overcoming the challenges that come alongside data.
What are some common sense and easy things companies can do to improve their data approaches?
An easy way to start is always to think in terms of records or collections. Is it easy to access the historical data for a particular process? How long am I spending on ad hoc modelling finding out the answer to a client question?
A practical example: the data team works with employees in each department to detail the routine tasks or reports that clients expect to receive. From here, the team creates a Python software script that describes the steps to build the report from beginning to end. Once the output of the script is validated by the department, the data team can quickly optimise and integrate it into the production data pipeline.
Usually, the initial code is written in a Jupyter Notebook which is then enhanced to serve as documentation for a particular process. This can also be used later as training material during the onboarding process for new employees.
Finally, the user receives the data in the desired format, or it is integrated into a BI dashboard. The benefits of this process are that the level of technical expertise is low, it gives the engineers a longer runway to build an integrated and more robust solution, and it provides the department with an automated process straight away.
What does the future hold for AM RE relating to getting the most out of data?
In 2021 our roadmap will be focused on optimizing the client experience for our application, the AmDex, which is our in-house proprietary software. Through this application, we are automating the multi-step cross-departmental processes to make AmDex a one-stop solution that integrates all departments (underwriting, analytics, legal, and finance) into one centralized database.
AmDex creates transparency, reporting lines, and powerful visuals. It updates in real time and is central to the operational efficiency of AM RE Syndicate. Essentially, this will open our internal system to clients and give them access to data related to their portfolio.
Thank you for the recommendation and useful information. In turn, I would also like to share a real find. I recently came across college studying images and was surprised by the quality they offer. I recommend you to join them!