If McKinsey has done their due diligence, the global insurance industry is going to look very different by 2030. By their estimates, the continued introduction of new technology like the internet of things (IoT) and artificial intelligence will radically change the way that most insurance providers do business—paving the way for smart, automated workflows that reduce much of the need for paperwork and manual interventions. As a result of these changes, McKinsey estimates that fully 25% of positions in the industry could be automated or consolidated by 2025, and that by 2030 the number of personnel associated with claims in particular could be reduced by more than 70%.
When this happens, application testing within the insurance industry will be more important than ever. In fact, if you’re working in insurance right now, you’re probably already noticing that more and more of the process of setting and enacting policies, assessing and paying out claims, and billing/invoicing have gone digital. This means that the smooth functioning of your software is more important than ever—both internally and externally.
If you’re in this position, there’s a good chance you’ve encountered the idea of test automation as a potential path forward for increasing efficiency and managing digital transformations. Perhaps you’ve even done some online digging to learn more about it. But what is test automation, and what impact can it have on the modern insurance industry?
What Is Test Automation?
Okay, let’s start with the basics: within an insurance context, test automation is exactly what it sounds like—automatically verifying that your digital technology (whether that’s a client facing app for filing claims or internal invoicing systems) works the way it’s intended to. Because so much of what’s done in insurance is gathering data in order to set policy rates and assess claims, most insurance providers field-testing their applications would likely benefit from automating end-to-end tests—i.e. tests in which an action is carried out at one end-point (say, a user’s phone), and the test checks to make sure the action has had the expected outcome at the other end-point (a claims adjustor’s work laptop).
Traditionally, this kind of process would have necessitated an individual user signing in to an end-user account, performing the required action, and then signing in to the corporate-side account to ensure that the information (the date of the claim, say) had shown up in the correct place. The user or users would then repeat this process for every distinct action that required verification, which could quickly become time consuming. Automation speeds this up by cutting out the middle man (i.e. the human), and running through test cases automatically. Depending on the automation provider, this might involve using something like Robot Framework to create modular, keyword-driven test flows that can be adapted to changing technological needs fairly easily.
Why Test Automation?
Like we said above, as digital technology becomes more and more prevalent in the insurance industry, insurance providers will have to devote more and more resources to service verification for that technology. Thus, at the simplest possible level, anything that can be done to make service verification more efficient will increasingly have the power to drive positive ROI. To get a little bit deeper into the weeds, this transition period (in which the industry is evolving towards increasing levels of digitization) is going to have particularly stringent testing needs. Why? Because this is the time when businesses are most likely to need to migrate entire systems, attempt to integrate legacy technology and workflows, and incorporate new technology like the IoT into existing workflows. Needless to say, this all has the potential to be messy.
Because the first steps in test automation involve a) identifying the areas that can be most effectively automated (typically rote and repetitive tasks), and b) scripting up the use cases for your automation framework to test, insurance companies can think of test automation not just in terms of time savings, but in terms of future-proofing. This is to say that automation encourages you to map out your software ecosystem in a legible way, and that that fact has the potential to offer operational benefits above and beyond improved time-to-market. As you design more and more use cases for automation for your various pieces of technology, you not only speed up future tests, you also speed up future design and implementation processes. How? By creating ever more documentation and data that focuses on the ways that your technology is actually used.
Automated Testing in Practice
Okay, we’ve given a somewhat theoretical sketch of how testing automation and test automation solutions might pertain to insurance companies as they continue their digital transformations. But let’s take a moment to be a little bit more concrete and go over some of the ways that these workflows might play out in practice. Let’s say you’re rolling out a new app that tracks automotive usage in real-time for car insurance policy holders subscribed to a pay-as-you-drive model:
- First, you’ll have to identify all of the use cases that require testing; in this case, that might mean identifying all of the different types of data that the sensors could transit (mileage, gas usage, etc.) and making sure that they transmit the right data to the right place at the right time.
- Once you’ve sketched out these use cases, your automation framework will perform the tests on devices (in this case the various IoT sensors that would appear in the vehicles) for each of the use cases.
- Depending on your solution, the automated tests will potentially run at a rate of hundreds of use cases per day—much faster than a typical manual end-to-end test could be conducted.
- If you’re using a keyword-based testing framework, your solution will provide readable documentation based on your pre-defined keywords. These will help you to pinpoint and address any bugs, while also preparing your operation for future tests.
From here, the move is simply to rinse and repeat. Most businesses tend to start out small, automating tests for one function or area in particular (like the auto sensors), before expanding into other areas (whether that’s creating and renewing policies, assessing claims, or what-have-you), slowly decreasing their manual testing efforts over time. In this way, you pave the way for easy regression tests and a culture of testing—both of which reduce bugs and service gaps in the long run.