Can Computer Vision do wonders for the insurance industry?
We are living in a world of autonomous vehicles, health tracking sensors, and where a plant can talk to a human. A new, interconnected world that requires proper risk analysis strategies.
I am not an insurance specialist, but I work as part of an innovation team at AdvantageGo where we build technologies for the future. We understand the needs of insurance clients and the current market to deliver innovative products that allow insurers to create an intelligent digital strategy. As I see it, underwriting excellence and claims processing are the keys that underpin the insurance industry, and in my humble opinion, emerging technologies will fuel the need for new ways to analyse risk, new risks and mitigate claims.
What Is Computer Vision?
Computer Vision, as the name suggests, is about a machine having the capability to identify objects in their surroundings and take actions accordingly, just like us. Humans have the capability to see the beauty of nature, our neurons helping us identify and interpret objects. The rise in innovative hardware technology makes it possible for machines to do the same.
Computer Vision unveils the context beyond image recognition and understands the relationships between objects. Previously, computers could ‘see’ a few people in a room with plates and a frying pan. Now, it can recognise that the image it ‘sees’ is about adults cooking dinner together. ‘Sensing’ technology maps objects and their location and interprets how objects are positioned within the real world. For example, autonomous cars use ‘sensing’ technology that identifies the position of traffic lights, zebra crossings and pedestrians – the technology analyses all this data to adjust the car’s response per its surroundings.
How Computer Vision Can Disrupt the Insurance Market
How can we apply Computer Vision within the insurance sector? I’ve outlined a few case scenarios.
Property and Casualty Business
Under property and casualty business, the process of inspection and claim settlement is time-consuming when factoring in the time it takes to analyse and assess the damage to a building or car. In this scenario, Computer Vision has the potential to significantly speed up the process, reduce errors, and lower fraud. Via satellite images, drones and big data, computer-assisted inspections are now possible.
Estimating a loss is still somewhat challenging for machines to predict, but there are some situations where relatively simple geometry is used to estimate a loss. Damage to a pre-fabricated home is one of the use cases we can consider as these types of homes usually have a simple layout and are generally fabricated using a set amount of building materials. Upgrades in technology enable computers to identify not only the object but its size as well. After obtaining such predictions, a machine can also predict the cost of replacement.
Drones are increasingly being used to perform damage inspections. Some insurance companies are using them to not only perform identification and classification but also provide the added value of reducing the risk of harm to adjusters. As an example, an inspection of damage to a rooftop can be dangerous to the adjuster who must physically assess the damage. A drone can easily capture detailed images of the roof, including parts of the structure that are difficult to access.
Artificial Intelligence (AI) powered image analysis, stitching images, object identification, and analysis bring rooftop damage analysis for insurers to the next level.
Is this the only use case?
What I’ve outlined so far is just the tip of the iceberg regarding Computer Vision capabilities. It can achieve more than just helping to identify and adjust claims.
If we look at risk prevention or risk mitigation, Computer Vision can be used to achieve the same. Autonomous vehicles need real-time object detection, which leads to collision avoidance and helps to prevent claims from ever happening. Collision avoidance is risk management taken to an entirely new level of sophistication.
Implementing new technologies such as AI or robotics is only one piece of the puzzle. With good analytical skills and business expertise, insurers can take advantage of emerging technologies with inherent risk knowledge to forge new ways of underwriting risk.