By Harrison Jones, ASA, Actuarial and Insurance Solutions at Deloitte, and Bernice Lim, FCIA, Actuarial Consulting at Oliver Wyman
Most actuaries work with data and build models, but do most know how to facilitate the distribution of their data and models? Sure, email and shared folders can perform the task, but how do you maintain an appropriate level of governance and standardization around what is being shared? The purpose of this article is to introduce the idea of application programming interfaces (APIs) to actuaries who may not have considered them as a way to distribute their data and models.
What is an API?
An API is a connection point that allows different computer or software applications to interact with one another. Such interactions are not limited to the internal systems of a company as they permit communication among internet-enabled computers.
Through APIs, a program or server can control access to its resources by making certain parts externally accessible while keeping others hidden. APIs come in different categories (e.g., SOAP, REST) and each category has its own protocols and architecture. These protocols define the rules and format of how the API operates, which should be considered when selecting the appropriate API protocols.
Why are we interested in APIs? APIs can be applied to different aspects of a business. Of relevance to actuaries, an API allows users to access and extract data from software that can be useful for actuarial analysis, expand on the functionality of an actuarial model while leveraging other internal or external software, and integrate different systems within a company to achieve maximum potential and automate processes to improve productivity.
How do insurers use APIs?
APIs can add value in several areas of an insurance company in line with the digital transformation efforts that are critical in today’s world. We will explore several common uses of APIs among insurers.
Insurance company operations
An insurer can have its own APIs to streamline its operations and connect different tools within its enterprise. Consistency across systems through APIs allows for easier deployment of new systems or models. It also encourages collaboration within a company to generate higher business value; for example, by allowing for quicker and wider access to company data for an analytics team.
Insurance claims processing can be more efficient when customers file using a digital claims system that is integrated with an insurer’s other systems through APIs. For example, in the auto insurance industry, claims processing can be accelerated by allowing policyholders to upload claims photos and obtain approval quicker than via traditional paper claims.
Marketing or sales
APIs offer opportunities to expand marketing beyond traditional channels and improve insurance application processes. Insurers can partner with third parties to offer their products on external platforms. For instance, Lemonade, a digital insurance company, launched a public API which allows partners to offer Lemonade insurance policies through third-party websites and applications. In addition, APIs can streamline different stages of the policy application process, including application submission, premium quoting, and underwriting.
How do actuaries use APIs?
For actuaries in traditional roles, their use of APIs were often limited to defining the API requirements. A developer would take those requirements and build the API for the actuary. Subsequent use would typically be limited to either
- updating/managing the technical model which sits at an API endpoint, or
- receiving data from an API request (the request being built by another team)
However, new possibilities have emerged as the role of an actuary transitions to being more tech-savvy and as actuaries become better versed in programming languages. Instead of an actuary simply defining API requirements, they now have the capacity to build the API components themselves. The following examples of actuarial use cases, which were traditionally constrained by skill set or system limitations, could be enhanced by APIs, the common denominator being that APIs facilitate faster and more reliable communication between different systems:
- an advanced rating algorithm for a personal auto insurer;
- an underwriting classification algorithm for a life or health insurer; and
- an algorithm that assists in detecting fraudulent claims for an insurer with a long-term disability product
What software is available to build and interact with APIs?
There are many options available to actuaries who want to expand their capabilities in building and interacting with APIs (the server-side development required in setting up an API is not addressed here. It is assumed that actuaries would rely on IT/infrastructure support teams to assist with this element of the API development life cycle):
Postman is a free platform for building and managing APIs. It is widely used; as of April 2022, Postman reports having more than 20 million registered users and 75,000 open APIs. Postman does have some features that simplify the process, but it’s been mostly used by developers who favour a code-heavy environment.
R (plumber)/Python (Flask)
R and Python are interesting options for actuaries who work in one of these programming languages. Both languages require intermediate to advanced knowledge – not a concern for actuaries who boast strong backgrounds in those areas.
Coherent, with their software Spark, is a no-code development platform for transforming Excel spreadsheets into APIs. Given the long-standing popularity of Excel among actuaries, Coherent Spark is a compelling option for actuaries who want to dive into the world of APIs but also want to continue using Excel for model development.
APIs have been used widely since the early 2000s, finding common use in the insurance industry. Actuaries have typically been, and will likely continue to be, on the periphery of API development, though. This isn’t necessarily a bad thing. API development takes a certain skill set to master, one well-suited to those with a computer science education (unless using a low- or no-code alternative like Coherent Spark).
That said, actuaries should recognize that APIs are often used to distribute production-ready versions of their data and models. Understanding API basics, even during the model-building stage, could go a long way to enhancing control and stability for the operations of an insurer.
This article reflects the opinion of the authors and does not represent a CIA statement.