Revisiting the 'Buy vs. Build' debate
Aug 2024
In P&C insurance, the conversation is shifting from a pure buy-or-build dilemma to a more nuanced question: how much to build and what to buy?
Aug 2024
)
The question of whether insurers should “buy or build” needs to evolve. It’s not a simple, binary choice like flipping a coin. Risk management is complex, further complicated by current economic pressures, rising reinsurance rates and the increasing frequency of severe weather. Many insurers simply don’t have the time or resources to build new tools. Thus, the conversation shifts from a pure buy-or-build dilemma to a more nuanced question: Is investing in external technology worthwhile? And to what extent should one buy or build?
Who buys and who builds?
In exploring who tends to buy or build, it’s clear that size and resource availability are decisive. For instance, historically, many insurers are more inclined to buy and adopt new technologies. Building a robust solution in-house requires significant time and resources. Even if an insurer could build their own solution, the software would need continuous updates to stay effective and efficient. Building any comprehensive software solution is difficult, and insurers already have enough on their plates.
Some carriers opt to build partial solutions, with an eventual goal of an end-to-end solution. If they have the infrastructure to build internally, they often will. They may even possess data science and engineering teams dedicated to creating a customized solution, and building in-house allows them to maintain control over their data and processes. However, this approach still presents significant challenges. Despite available resources, it’s nearly impossible to avoid long development timelines and continuing maintenance needs while keeping up with the latest tech advancements. Building a full end-to-end solution is an arduous journey.
Although the build option varies for each insurance carrier, it is a huge challenge.
So, should insurers simply buy an external solution? Not necessarily, as not all solutions are created equal. The key is finding the right partner and unlocking the most value.
Strategic collaboration
When selecting an external technology provider, insurers should understand the provider’s experience, capabilities and the comprehensiveness of their solutions.
Here are a few key aspects to look for:
- Transparency is paramount, particularly with AI technologies. Insurers must ensure these technologies are explainable, to minimize errors and biases, aiding rather than complicating decision-making processes.
- Adaptability is crucial. The technology must be flexible enough to meet specific insurance needs and accommodate regional differences. For instance, insurers in Washington and Louisiana may both want insights into property conditions across their portfolio but require different predictive risk scoring for wildfires or hurricanes.
- Testing is essential. Insurers should demand demonstrations of the technology’s effectiveness in real-world scenarios, such as the ability to remotely monitor roof condition for staining, ponding or rust.
A truly innovative solution should also provide comprehensive support and continuous improvement to ensure the technology evolves alongside the insurer’s needs. If these criteria are met, insurers are much better equipped to reduce risk, predict and prevent losses and improve their expense and loss ratios.
The importance of customizability
Suppose a Northeastern insurance carrier relies on a small underwriting team to assess 50,000 properties across their portfolio. There’s simply not enough time or resources to manually inspect each property.
But if they leverage a solution with high-res and recent imagery paired with accurate computer vision models, the system could accurately identify which properties are at risk and which are not. If the system can automate and straight-through-process low-risk properties while flagging high-risk properties, the underwriting team can move much faster and focus their attention where it’s needed most.
Consider the same issue for a better-resourced insurance carrier. Perhaps they’ve already developed some front-end functionality in-house to monitor property risk. Now, suppose they want specific AI models from a vendor trained with high accuracy for dozens of other property attributes. What should they do? In that case, the right vendor could seamlessly integrate their models with the insurer’s existing systems to provide a tailored solution.
Perhaps that same carrier has solved many of its underwriting challenges. Instead, they may need to buy the back-end—like a catastrophic weather or disaster response system—to help manage claims and the impact of severe weather. Again, the right partner could provide access to the latest cutting-edge technology, set up proper workflows and eliminate the burden of development and maintenance.
This customizable approach allows larger insurers to balance proprietary innovation with leveraging market-proven technologies. Think of it this way: If you’re making a cake, you might have all the baking tools, but if you’re missing eggs, you wouldn’t buy a whole farm just to get eggs. You’d simply purchase the eggs you need. Similarly, if you lack all the necessary ingredients, you’d buy a ready-made cake. This principle applies to insurance technology, as well: Rather than investing heavily to build everything in-house, you buy the full or partial solutions based on your individual needs.
Final thoughts
The “buy or build” debate in P&C insurance is nuanced, heavily depending on each insurer’s unique circumstances, and will likely continue to evolve. So, it’s important to be a resource to help insurers evaluate their options. Strategically choosing when to buy and when to build enables them to leverage the best of both worlds, ensuring competitiveness and efficiency in an evolving industry.
Reprinted with permission from the July 14, 2024 edition of Insurance Thought Leadership. ©2024 Insurance Thought Leadership, Inc. All Rights Reserved. Further duplication without permission is prohibited. All rights reserved. Original article can be found online here.