We've told you before, including in our last magazine and Thierry's blog, about our new perspective on data. A change that has taken leaps and bounds since our partnership with Hippoline, the arrival of ANVA Hub and switch to a new data platform. Team Data within ANVA will soon celebrate - in addition to its one-year anniversary - the kick off of the very first pilot!
Product Owner Marline van Ramselaar starts, "Our customers are increasingly looking at how they can use data inside and outside the ANVA environment. Think of tools used for business steering, the interpretation of duty of care and AI software used in the claims and underwriting process. We also see customers developing applications (or having them developed) on top of ANVA's data. All data extraction and queries in ANVA 4/5 now happen directly on customers' production environment. This makes ANVA slow, as a lot of processing has to be done in the background. That works anything but ideal for our customers. By deploying our own Operational Data Store per customer (Dataplatform), we offer one environment where all data for each customer comes together and can be accessed individually."
A Data Platform is a collection of data put in such a form that recurring and ad hoc queries can be answered in a relatively short period of time, without putting undue strain on the source systems themselves. The move to the DWH is a very big step for and within ANVA, but also ensures that we are making our software more robust. Deploying the Data Platform will allow us to add ANVA Hub data and external sources in the future.
The deployment of a Data Platform for data extraction arose in 2021, right after the official start of our collaboration with Hippoline. After the successful introduction of the BI dashboard called HIP (Hippoline Insurance Platform - of which the DWH is a crucial part), we were soon convinced of this solution. In addition, Team Data saw opportunities to replace our ODBC connector, which pulls data directly from ANVA 4/5, in order to significantly reduce system load.
"The Data Platform also includes the ODBC connector which is similar to the old ANVA 4/5 ODBC connector. So we started a technical study to see if this could be a good replacement. Initially, we worked with Hippoline to ensure that all data was structured and modeled for HIP. We then set up the DWH so that we can start using it for HIP as well as for the new ODBC link. Now that that's in place, we can start the first pilot," Marline says enthusiastically.
In August, a number of ANVA customers will start testing the DWH, each with their own ODBC connector. Marline: "We will let them test the extent to which they can work with it. For example, we will investigate what happens when they first address connector A and then replace it with connector B, which is linked to ANVA from the Data Platform. We also want to investigate extensively whether the new ODBC connector meets everyone's needs, whether all the necessary files are available and at what interval new files should be made available."
Besides the various tests at and by ANVA users, team Data - in collaboration with ANVA partners - will also conduct several tests of its own. In this way, we hope to gradually find out what the working method using the Data platform means in terms of data extraction, storage and costs(compute and computing power). "The results from this research will form the basis for the eventual standard packages and pricing models. For this reason, we very deliberately choose very different types of pilot customers. From large to small, proxy and provincial," explains Marline.
At the end of the summer, when the pilot is complete, Marline's team will process its retrieved results and fine tune them where possible. Then this fall, they will start transitioning ANVA customers to the Data Platform in batches. Of course, a hybrid situation will remain available for the first period. "The old ODBC connector can still be addressed in that case, but also the new link from the Data Platform is already available. Then we will make the full transition to our new, future situation: working from the Data Platform."
