Chubb builds IoT stacks to pinpoint risks and underwriting

Chubb finds value in IoT technology to prevent losses

November 2, 2022 6:30 AM

Chubb headquarters in Whitehouse Station, New Jersey

Digital Insurance spoke with Hemant Sarma, senior vice president and global digital Internet of Things lead at Chubb. The global carrier needed a good reason to acquire a water leak detection technology company as it did with StreamLabs last year. Sarma spoke about Chubb's business case for bringing an IoT technology into its corporate fold, and how it reduces risks and lowers premiums.

What is Chubb trying to accomplish with IoT?

Hemant Sarma - Chubb.jpg — Hemant Sarma, senior vice president and global digital Internet of Things (IoT) lead at Chubb

My role is to find solutions that apply to Chubb particularly. We find a lot of vendors out there that have a solution looking for a problem, and we may not have the problem. My team sees how a solution can be adopted technically, makes a proof of concept or program around that and then embeds that program into our ongoing insurance offering.

For example, water losses, non-weather water damage losses, losses that emanate from a non-flood event, like a pipe break. That loss in the industry is more than $20 billion. Chubb is a big share of that, because we write a lot of this business.

We use two types of technology. One is sensors that detect water or temperature drops that can cause a pipe break. Another piece is from StreamLabs, the company that we acquired [in August 2021], which has an automatic shutoff feature on a device that is embedded in the water line in a single family home. When it detects an unusual flow, it shuts the one line off within seconds, preventing any further water leakage in the home. That prevents losses from happening, for which we might have paid millions of dollars.

What’s the IoT foundation for using a technology like StreamLabs?

It's what we call the IoT stack. The IoT stack is the devices, the back-end communication, and a visualization tool, whether it's in an app or a dashboard. We completed an entire stack with the [StreamLabs] acquisition. Why would Chubb get into an OEM [original equipment manufacturer] company, being an insurance carrier? The benefit for us is that we now have insight into how to price a product, which is typically a commodity. How do we embed that product into our insurance offering, and build what we would consider a long term solution? Our ultimate goal is to prevent the loss. It's not about generating revenue on the front end from these devices, it's more about saving losses for our customers.

It also helps us as an insurance company think about this. When you do the underwriting, your touchpoint is once a year. When you do risk engineering, it is maybe two or three times a year. With the sensors and alerts, it's seconds. Every second you are getting the alert, there's an opportunity to reach back to the customer for whom you're providing an insurance solution to check on them.

This also increases the value that the customer sees from an insurance carrier, which increases retention. That is a benefit of IoT that I see coming into play. IoT touches every single aspect of the insurance cycle. If you reduce losses for the customer over time, the risk profile of the customer improves. If it improves, their premium's going to be reduced over time. They get a lower deductible.

In this example, how does the use of IoT affect underwriting?

With the new data that's coming in, you can better underwrite the risk. You can also create better pricing mechanisms. In the early days of the 1900s, there were no sprinklers. When sprinklers came on board, then you had buildings with sprinklers getting a lower rate on coverage. So you have two rates -- one for a building with sprinklers, another for a building without sprinklers.

The vision is that as more of these sensors are embedded in buildings, hospitals and colleges and everywhere else, you would have a separate rate for a water mitigated, water-sensored building versus a non-sensored building. That's going to be powerful for any carrier that can do that. You'll have a better segmentation of the price.

Is that rate discount in place or in the future?

It's not there yet, but what we are doing is that we are providing incentives for the customer to build that. On the consumer side, the personal lines side, we actually pay for those sensors. Chubb pays for it. It's embedded into our contract. On the commercial side, we are providing incentives for them to adopt. But it's not in the ratio that you would then come up with the pricing. It's still way off. Right now, to your question, that is still the future.

It's still being worked out because the challenge is getting to the actuarial level.

What other coverages benefit from IoT, and how?

We are experimenting with camera-based AI for slip and fall. When you go to a mall in the U.S., when you walk in, there's always the potential for slip and fall. A lot of accidents happen. People slip. In a workplace, it could be a worker who gets injured at the site. Can you influence their behavior? Without getting into areas of privacy, data security and compliance, can you actually improve the behavior of the customer or the employee or worker, by having other tools available that can reduce the loss? It's all about loss reduction, so we are experimenting with that.

Another use case, in every large building, you have the boiler or the chiller that runs the building -- the air handler. Every building has some of those components. Equipment breakdown is a big factor. If it breaks down, you don't have power or you don't have heat and water. Because of supply chain issues, a lot of these replacement parts take months and months. With predictive maintenance, with sensors running, you can prevent or give a heads up.

Remember driving old cars? You listen to the vibration of the internal combustion engine. It vibrates in a different way when there's something wrong with it. The vibration cadence of a machine can actually tell you that something's wrong with it. There's not enough oil in it or there's not enough lubrication.

Every machinery has a certain cadence. Vibration makes a certain noise when it's okay, when it's lubricated, when it's not. So we can pick those sounds and vibrations. Let's say it takes two hours to heat up a large commercial boiler. Over time because of scaling, it is taking three hours, four hours, five hours or more. You do know that there's something going on inside it. You could use some of these KPIs to identify that there's something wrong with the boiler. It becomes a predictive idea of how you can think about it.