The commercial insurance market is slow and fragmented. While numerous platforms and technologies have been implemented in recent years to help digitise specific tasks and business lines, there is little to no connectivity between them, while many other tasks remain analogue. 

As a result, the process of buying and selling commercial insurance, which involves complexities in assessing risk, is still largely archaic. Brokers put in significant effort to deliver data to insurers, who then must thoroughly assess it, make a decision, and relay it back to the brokers.

Willis Towers Watson (WTW), one of the largest insurance advisory and brokerage firms, wanted to know if there was a way to digitise the whole process and create a digital marketplace where brokers can get information about risks to insurers, and insurers can return algorithmic decisions – or decisions in markets where algorithmic trading is not appropriate much faster. 

"[The question was] “can we effectively take something that traditionally takes hours, days and weeks, and turn it into something that takes seconds or minutes?” says Alex Morris-Tarry, the CIO tasked with turning this vision into a digital trading platform for the commercial insurance market, called Neuron.

“We're trying to move the industry from analogue to digital. We had to build a marketplace which brings insurers and brokers close together, makes it easier for them to work with each other, makes them more effective when they work with each other, reduces their costs and increases consistency so the data is more common.”

A complex marketplace

The premise of Neuron is relatively simple: a distribution channel and data pathway that digitally connects brokers with underwriters, enabling risks to flow at scale and with common data standards, across lines of business. However, achieving this would require a sophisticated and complex platform encompassing many different propositions, with a lot of data passing through.

Commercial insurance is a vast and varied market, ranging from small businesses requiring vehicle insurance to large enterprises needing to insure multiple buildings or cargo ships. Such variety means it can’t be assumed there is a single source of data or single sets of data that are valuable. The complexity doesn’t come from having very high throughputs like it might in other industries, but rather the need to be extensible enough to manage very high-value workloads. 

Choosing the right database system for the Neuron platform would be vital. Morris-Tarry outlines three key factors driving the decision, starting with the flexibility not only to deliver the marketplace’s API-driven connectivity, and the right workflows through the APIs, but also to get data to insurers in a way that meets their needs (Neuron also builds its own UIs on the APIs).

“We can't just put in a monolith and assume that's going to work for the next 20 years,” he says. “This is an internal system, it's a client-facing system and it’s a multi-tenant system. We're providing access to many, many different clients, both insurers and brokers. That means we have to be able to change while still giving access to, for example, older versions of APIs for people who can't move as quickly. So we need tools that give us the ability to flex and stretch.”

Security, naturally, was another consideration. In an industry which transacts on a lot of sensitive data, brokers and insurers alike must have complete confidence that the data they put in the system goes only to the places it's intended for. The final key consideration was ensuring that Neuron would have the ability to continually innovate and stay ahead of the game. 

“We want to make sure the tools we use let us continue to innovate and build our solution in a modular way,” Morris-Tarry adds. “We use best-of-breed technologies and as new opportunities emerge we want to be able to adopt them and compose the solution that produces the best outcomes for our customers. Our ability to move with the times is one of the major selling points.”

Thinking about the future

Neuron’s need for flexibility, in particular, whittled down the options quickly, as a traditional relational database was not the answer. Morris-Tarry knew the capabilities required to deliver Neuron would lie in a document model, and MongoDB ticked the boxes of scalability and security while also providing assurance it would continually evolve with the latest innovations.

“MongoDB has the flexibility we need without having to create a hugely complex, difficult to maintain, difficult to update… set of relational databases,” says Morris-Tarry. “You can build the data in the shape you want to build it, store it in the way you want to store it and change it in the way you need to change it. The flexibility of MongoDB Atlas really shouldn't be underestimated.

See also: "Everyone was laughing: Now they take us more seriously..." Europe’s biggest retailer turns cloud provider

“If you look at what MongoDB talk about you can see that they care about new technologies and the use cases they will open up. Atlas Vector Search is one example and they've also announced the Application Modernization Factory. From my perspective, that's not something MongoDB needed to do, but it demonstrates they are thinking about the future of technology.”

The major barrier to overcome when deploying MongoDB Atlas was one of mindset. Neuron’s IT team, like many others in the industry, initially approached databases with a relational mindset, which can limit flexibility and slow down systems when using the document model. However, with the support of MongoDB’s professional services, they quickly adapted to a new way of working — unlocking greater agility and performance.

“[The mindset shift] was the first step, but we got there and the outcome was an incredibly fast system managing our data and working with the other components of our solution. Now, our environment sits with MongoDB in the centre. It’s our version of the truth. It's our document store. We can analyse data that's going through the system easily with the additional Atlas tools. 

“If you think you need help in doing some of these implementations, reach out to MongoDB because our experience is they really know what they're talking about. You get a high-quality of support and direction at a reasonable price. For example, with the dashboarding that Atlas provides, we were able to get some support on building that out really quickly from MongoDB.

"Now we have an analytics dashboard that we can use both internally and for our customers as well, so they can see how their data is performing against the information coming through the system.”

The result? A platform that maximises insurers' risk capacity, brokers' access to capital, and which creates robust, flexible data foundation for evaluating risks, with the ability to augment data packets from third party sources, and which minimises low-value, duplicative work across the ecosystem.

Delivered in partnership with MongoDB

The link has been copied!