Fintech Insights: IBM & ING Building Better Data Fabric for Financial Services

From data lakes and warehouses to on-prem servers, accessing data has never been more complicated. A financial services organization or any fintech company could benefit from data fabric. it could be the answer to complex challenges related to Big Data management and governance.

IBM believes the answer lies in data fabric, which can discover and unify disparate data sources – be it a data lake, catalog, or warehouse. Bringing them into one unified view can give users a single point of access to find, understand, shape, and utilize data throughout the organization.

At THINK 2021, Ferd Scheepers, Chief Architect of ING Tech Group Services, talked about how the firm has partnered with IBM to develop his dream of data fabric – AI technology now part of IBM’s Cloud Pak for Data.

ING has been building vast data lakes to store its data – but has struggled to make them work. Why is that now a problem? 

Fred: There are essentially two issues. One is the amount of manual work that it takes to bring governed data into our data lake, which has to involve a lot of subject matter experts. It’s quite a challenge and cost to do this, and it is becoming an inhibitor of getting more data into the database.

How Artificial Intelligence Is Revolutionizing Data Capture

The second one is a more recent issue.

More and more data is moving to the cloud, applications in the cloud pull in more data, more data pulls in more applications and it’s a never-ending story. I need to govern my data in the cloud, in the exact same way as I need to do that on-prem, but the quality of the data and the plethora of different tools that I have in the cloud makes it very, very difficult for me to govern my data. And that’s given me quite some headaches.

You’ve obviously spent a lot of time thinking about the issue– what do you think the solution is?

Fred: When I was thinking about this problem a number of years ago, when we first started talking to IBM, I suggested a kind of abstraction layer between the data and the different consumers of the data across multiple clouds. It should take away the complexity that we now have moving data around and understanding data in different environments.

Ideally, automation should make it possible for me to just pull in data from the different sources, understand that data, and use AI to start mapping it directly to our company language (ING developed its own global glossary called Esperanto that describes the business terms it uses frequently to keep the language consistent).

This sounds like a huge undertaking – how is ING planning on managing the shift?

Fred: You can’t just drop it in there and say ‘hey, we’ve invested in a central data lake for eight years and now we’re going to do something completely different’. So the idea is that we slowly move our current setup to AI data fabric.

We’ve already done a lot of the work by classifying the data in the IBM toolset, this will be preserved, and we know the data, and we’re going to slowly move from what we have today to a mesh by first implementing this on-prem, and then extending it to the different clouds where we also bring data.

What are the biggest governance issues you’re facing with this move to data fabric?

Fred: I think there is a need for keeping governance very central. You really want to have that central location where you implement your company language, where you set your policies, and where your data owners and also the regulators can see you are in control of the data.

In an industry as heavily regulated as banking, this is essential. We need to be able to prove to our regulators that you can only have access to data on a need-to-know basis, especially for sensitive customer data.

It’s also critical to be able to show regulators that you know that no matter where that data is consumed, you know who consumed what data for which purpose at what point in time, and that everything is done according to the policies that you’ve created. All this should be described, ideally, in natural language, which then translates into real rules that need to be executed at the point of enforcement. That’s the aim.

How much do you see the data fabric performing this orchestration role?

Fred: It’s my dream that this is hidden away from the consumers of the data, that’s what we need to aim for.

I see this as an abstraction layer, and a little bit of AI that understands what is possible, based on a lot of strong policies, I really want data fabric to handle all these things and do 100% of the heavy lifting. I just want to consume data. That’s what we need to aim for.

You’ve been thinking about this a lot, what is your advice to other large enterprises thinking about moving to data fabric?

Fred: The first thing is really don’t wait – there’s a lot of stuff you can already do today. Make sure you’re bringing your data under governance, and make sure you really understand your data, and start planning for the journey ahead.

My second piece of advice would be to engage with a trusted partner. This is an industry that’s moving a lot, and you can assemble your own data fabric, but then you’re doing the heavy lifting yourself.

At ING we’ve been there, done that – and don’t really want to do that again!

When we work with a partner like IBM, where I know we can build something together that will work, and take that heavy lifting away.

Finally, start talking to people within the organization, telling them what’s coming, that this will facilitate the journey to putting all your data in the cloud.

[To share your insights with us, please write to sghosh@martechseries.com]