Agile insurance pricing: Rapid reactions to dynamic market events

Dawid Kopczyk
Dawid Kopczyk
Chief Executive Officer
August 28, 2023

Over the last decade, the insurance industry witnessed tremendous changes including customer digitalisation, setting up higher expectations regarding insurance product distribution as well as a surge in new tech enablers. The market has become much more dynamic in response to rapidly changing conditions.  

Nowadays, pricing and underwriting teams experience many market events such as inflation, new data, the ever-changing regulatory landscape and competitors’ movement. Particularly in P&C business lines, we can observe a constant pressure to deploy new pricing changes to avoid losing portfolio or face increased loss ratios. The ability to successfully compete in such an environment has never been more important. However, according to McKinsey's report [1], pricing and underwriting teams are facing numerous obstacles when reacting to such market events, mostly connected to having fragmented pricing systems based on legacy technology.  

This has been also proved by the MunichRe’s report [2], where according to their survey 47% of respondents desired to improve the way in which pricing updates are deployed to the sales channels.

A group of rectangular bars with textDescription automatically generated
Figure 1 – Results of MunichRe survey with regards to pricing deployment needs of insurance companies. Source: [2].

In this post, we will dig deeper into the roots of issues with pricing deployment as well as present considerations concerning how to improve it.  

Roots of Deployment Evil  

Being able to create a successful pricing model is just a part of the story. What is also crucial is being able to deploy it to the sales channels so that the benefits of what was built can be achieved in the production.

Over the discussions with our customers, we have identified 3 sources of issues:

  1. Dependency on the IT to deploy pricing updates
  1. Fragmented pricing architecture  
  1. High cost and effort for pricing updates’ deployment

Dependency on the IT to deploy pricing updates

Once an insurer decides to implement an in-house solution allowing to deploy pricing strategies to the sales channels, the project usually takes several months to complete with specific requirements given upfront to the IT team. However, the reality of insurance market dynamics makes these requirements quickly outdated and incomplete. This in turn limits pricing departments and makes it difficult for them to maintain the desired level of competitiveness. As an example, the price walking regulation in the UK caused a lot of implementation headaches for insurers' executives to successfully embed it in their pricing deployment processes.

Fragmented pricing architecture  

Switching between different solutions for data processing, modelling, reporting, deployment and monitoring causes a significant increase in time-to-market of pricing updates, or even makes it impossible to deploy some of them due to the non-compatibility of different parts of the legacy technology. For instance, online price optimisation cannot be deployed to a simple rating engine since it cannot be expressed in a set of rating tables without loss of accuracy.  

High cost and effort for pricing updates deployment

Having a fragmented pricing process without automation, governance and monitoring put in place leads to the high cost of a single pricing update. Reacting to events such as competitor pricing movements might require numerous A/B testing and updates of pricing models, including implementation of more sophisticated modelling techniques, personalisation, behavioural models, or integration with external data sources, just to name a few. These might be extremely resource consuming if we don’t have a proper technology implemented.  

New Hope  


A typical insurance pricing process involves many steps from data ingestion, through data processing, modelling, setting quoting rules, price optimisation, and finally deployment to the sales channels with the ability to monitor it in real-time.    

A screenshot of a computerDescription automatically generated
Figure 2 – Typical pricing process for personal lines.  

Here is the recipe to avoid headaches during the deployment of pricing models:

  1. The interconnectivity and compatibility between the analytical part where we design a pricing model and the part in which we deploy it to sales channels is extremely important. The reduction of dependency on the IT is vital.  
  1. The same applies to the analytical part – passing datasets, models and reports between different parts of the pricing process must allow for great flexibility, while at the same time assuring proper governance of the whole pipeline. This allows us to avoid a fragmented pricing process.  
  1. Proper technology to perform deployment, perform real-time monitoring and A/B/X testing of deployed API is crucial for the early discovery of problems within certain segments of our portfolio and reducing efforts of recognising triggering events and implementing a new pricing update.  

Pricing actuaries and underwriters should build the processes in a way that assures these three rules are followed.

Figure 3 – Apart from real-time analysis and monitoring, Quantee 3.3 platform allows you to perform A/B/X testing on production. In this example, various versions of pricing models are exposed to the pricing engine traffic depending on sales channels and other variables.

Pricing Engine vs Rating Engine

There is a fundamental difference between pricing and rating engines with the latter restricted only to express a pricing model as a set of rating tables. Although convenient to implement that by the IT department or use built-in API in the core system provider, it is not enough to compete in nowadays market dynamics.

Pricing engines allow insurance companies to deploy any objects, including AI models, online price optimisations or complex pricing pipelines with nested models and data processing capabilities, such as external data source ingestions. They also unlock personalisation models to be included in a deployed pricing pipeline without any effort on the IT side. As a result, insurance companies with proper pricing engine technology can be agile and rise above the competition contrary to other players who are much more restricted by rating engines.  

Figure 4 – An example of a motor pricing pipeline with propensity-to-buy models allowing to return product bundles and enabling insurers to utilise full personalisation’s benefits.

Time-to-response considerations

Pricing engine APIs are stormed by multiple requests for a price from different sources, including but not limited to direct channel websites, quotation systems and aggregators. The APIs must be secure, stable, and most importantly respond within a reasonable amount of time which varies from tens to hundreds of milliseconds. This is even more important for the case of aggregators, who take into account time-to-response in their rankings of policy offerings.  

Figure 5 – Quotations from an exemplary aggregator in the UK (, which shows that not only a price is taken into account in the ranking but also other facts, such as response time.

A pricing actuary choosing the right technology for deploying pricing models must consider that topic and make sure that the solution is able to provide a response in a required time range, even for most complex pricing pipelines.  

Figure 6 – Monitoring of response times and health statuses of a complex pricing engine in Quantee 3.3 platform.

Your turn

Current insurance market dynamics in the P&C sector demands innovative solutions empowered by modern technology.  If you are in need for an end-to-end, integrated solution that is flexible enough to assure you will react to any market event and seamlessly deploy changes to the production environment, Quantee is here for you. We are always one step ahead of the game, and allow you to monitor, perform A/B/X testing, and real-time analysis in just one platform.  




Related content