The Future of Insurance Data, or the "BRAIN" Model

An insurance business is a living organism. One that is constantly evolving. But to do that, it needs a nervous system: the data infrastructure.

Because data irrigates and connects all departments. Consolidate it to build insurance products. Visualize it to understand. Analyze it to make the right decisions ... it's the lifeblood of the business. 

We won’t transform insurance without transforming our approach to data. That's why Seyna has set itself the mission of building the insurance data infrastructure of tomorrow.

We don't have THE solution. We're building A solution. This article is the first in a series of reflections written to engage concrete and collective discussions on the future of insurance data.

To get the ball rolling, let's look at what we consider to be the 5 pillars of the data infrastructure of the future, the engine behind the nervous system - the "BRAIN" model: 

• Business Oriented | Feeding your business as a whole
• Real-time | Uploading, distributing, and analyzing data in real-time 
• Accurate by design | Guaranteeing accurate data, right from integration
• Interoperable | Allowing integration with all operators, players, and solutions on the market
• Nodal | Compiling big sectorial data to improve our collective understanding of risk
  
  

Business-Oriented

Data isn't solely an IT matter. It concerns all departments and functions within the company and goes beyond the simple aspect of risk management.

It must permeate the very fabric of the business. Create a silo and you'll start seeing this part of operations clog up and slow down. 

At Seyna, we have decided to consider data as a business within the business. A structure with its own governance, an obligation to deliver results to all customers (internal and external), and continuous investment.

It has therefore naturally been placed at the heart of our operations to support all our use cases.

Today, we can generate all our ledger entries in just 1 click. Our scripts produce and check over a hundred reinsurance slips in less than 5 minutes. Our regulatory risk exposure calculations are automated. Our "Legal Bot" autonomously generates contractual documentation and warranty tables. Our audit team benefits from automatic analytic reports, etc. 

There's still a long way to go, but we'll always pursue the same objective: to free up our teams' time to focus on our customers and strategic issues.

Real-Time

This component is already a market reality. 43% of insurers worldwide claim to collect and analyze data in real-time to predict and assess risk. (Source: The Data-Powered Insurer, Capgemini) This is excellent news, as it enables us, as insurance professionals, to operate more proactively, and less reactively.

Creating ultra-contextualized offers, managing risk in real-time, guiding commercial offers towards coverages or segments with lower claims levels to improve the combined ratio, anticipating drifts, automating anti-fraud controls, etc., are all part of the process.  

Collectively, we have moved from a static to a dynamic vision of data. To get there, we all faced the same obstacle: data integration. 

And it is quite the challenge: how do you build a model that automatically integrates all insurance data - regardless of its structure? We'll come back to our approach in a later, more detailed article ... but in case you were interested in a little teaser… 

This paradigm shift promises not only greater customer satisfaction but also greater business continuity. 

‍

Accurate by design

"Garbage in, garbage out", aka the most structuring concepts in the world of data. But this leaves us with a key challenge: how to make it impossible to integrate erroneous data?

First, we focused on maximum standardization of the integration process. Using a palette of elementary transformation functions, we can adapt this process to any data pattern. We have also opted for a dual control strategy, which runs: i) on entry, and ii) continuously on stock. That’s why we set up a staging environment, to preview the database post-integration. This way, we can correct the situation before the data is even integrated. The aim is to avoid costly corrections emerging after 6 months of operations.

Using a dedicated table, we configure in just a few clicks the battery of tests to be run on a given database. Once validated, these are automatically applied to the stock, continuously.

‍

Interoperable

This brick is probably the one I feel most strongly about. 

We live in an economy of hyper-specialization. Companies are built to address fractions of the insurance value chain. If you can’t integrate with them, and pass on the benefits to your customers, you're shooting yourself in the foot. 

That's why we've made it a fundamental principle of our insurance platform. Our data infrastructure has been designed to enable integration and connect with any insurance organization (broker, insurer, reinsurer). Similarly, it enables us to integrate any third-party service.

Tariff production APIs with Coherent, electronic signatures with Yousign, payment management with Stripe... Seyna was largely built on its integrations. So we've made it a cornerstone of our value proposition. 

It's still a work in progress, but we'll make Seyna the platform that lets you integrate all the best solutions on the market, in just 1 click. 

‍

Nodal

Nodality is about convergence. Creating crossroads, pooling our knowledge.

Admittedly, this dimension is still a distant fantasy, but we are convinced that Open Data is the future of insurance. 

Indeed, the entire sector could benefit enormously from a large database open to its contributors. Public, aggregated, averaged, anonymized data... The aim is to give all incumbents access to all of the sector's knowledge.

It's a long-term vision, but we're already headed in that direction. Ultimately, we aim to enable all brokers to compare their business with the market. This will feed the progress of the insurance industry and enable us to create ever more relevant products.

This was an overview of the 5 pillars on which we are focusing our data development. We will detail each of these points in future articles. 

In the meantime, please share your feedback, comments, and disagreements with us, and let’s keep the discussion going!

‍