I’m Patrick Sergysels.
I'm responsible for the data management entity within
AG Insurance.
Our main objectives
are to set-up data governance within the company
and as well to manage data platform to allow our business
to exploit the data within the different use cases.
AG Insurance is a leading
insurance company in Belgium since about 200 years
now, we sell products on the different
type of insurances. Property and casualty, life
and as well group and healthcare products, for instance.
AG Insurance wants of course to remain a leader
within the insurance companies in Belgium
as we were for several years now.
And one of the key areas where we want to progress
in the future is to improve the user experience.
We want to be able to exploit all of the data
that we have on our customers to deliver them better services.
In the past, we had different platforms
to support classical analytics, B.I.
where on premise we had to analyze the data
from the contract and the claims that we were receiving.
And in parallel, we started trying to exploit
new types of data in Azure
in Date Lake, for instance,
data on pictures of claims and in videos.
to be able to do so, we wanted to move forward to a more modern
and integrated data platform in the cloud to be able
to really use those new types of data.
We knew as well that in the
modern environment
it is necessary to answer the customer much faster
than we did in the past. It's of course not a simple project to do,
but the benefits at the end are really worth it.
We did know that it was a challenging situation for us
because we had a platform, an existing platform
that was really suffering from several pains
and it was necessary to move forward.
We knew that the project would not be an easy one,
but at the end we get and see the benefits
that we can deliver to our customers
My name is Laurent Horion
I work for the data management team as data platform manager
and I am leading the track of migration to Azure.
The challenges that we faced on premise were multiple.
The first one was about scalability
to support the exponential growth of the volume of data
for our business.
The second one
was more about the elasticity that we were missing on premise.
In an insurance company, you have to manage closing periods.
Those financial periods are quite busy
and servers were struggling with the workload.
The third one was about some missing capabilities
to support advanced analytics.
The processing of images, videos
and audio files, for example.
In Azure, we can benefit from all those capabilities,
and the last version of that
the partnership with Microsoft started
from the very beginning of the project.
During the opportunity study we could define the architecture,
review it with Microsoft Architects.
We could benefit from Microsoft support
during the implementation of the proof of concept.
We could estimate the running cost of the platform
thanks to their support as well.
And finally, during the implementation
of the project, from the beginning until now, we could review
specific questions, specific issues with them on a regular basis.
The last aspect was the training path for our team members.
It was quite a change
we could put in place training paths.
with their support as well.
Hexaware knew our environment.
for a long time. They were in charge of new developments for AG Insurance
and they were as well taking care of the monitoring of our platform.
As we knew they were Microsoft
certified and partner,
they had technologies to facilitate migrations to the cloud,
we decided to collaborate with them.
When migrated to Azure, Hexaware continues
to take care of the monitoring of the platform
and of new developments.
A few years ago,
we started by selecting Microsoft Azure environment itself.
We were already familiar with some Microsoft technologies,
like Power BI, SSIS, SQL Server.
The technologies we selected in Azure
for the storage are on one side ADLS Gen2
to store raw data, to store as well
videos, photos, audio files and so on.
Unstructured data.
We selected Synapse to store our data marts
to publish more prepared data as well.
On top of it, we decided to continue to use Power BI
to provide self-servicing capabilities to our business stakeholders.
Visualization and so on.
We used Databricks to implement our framework
on top of it, we decided to implement those technical services
to improve the automation of IT activities,
the standardization of it, for example.
And then we implemented
our data pipelines using data factory.
The first results of the production environment are quite encouraging.
The first one is about the total cost of ownership
that decreased by about 25%.
Another one is about the operations
that are running 4 to 5 times faster than before.
Then we
succeeded to reduce the cost to implement new projects
thanks to more automation, for example,
and we decreased the time to market
to deliver those projects to our business in some cases
we use as well sign up Serverless to get data
from Microsoft Dynamics 365
and integrate it in our data platform.
The last component of the system is the end user zone.
That's an environment where business users
can like, for example, a data scientist.
The analyst can implement their models, can query their data
and further prepare their report.