Silverlake focused on driving AI adoption

“There's no point to ‘meng-AI-kan' your bank if you're only going to do it on a silo-by-silo basis. Then you can't get the returns on your investments,” says Cassandra Goh, Group CEO of Silverlake.

Having established itself for more than three decades, Silverlake has been an integral partner for many tech companies in the Southeast Asia region. Over the years, Silverlake has grown to partner the world’s leading tech vendors to drive transformation in the banking sector.

Throughout its history, the service provider has gone through several changes, from acquisitions to new partnerships. While the focus remains on the financial services industry, Silverlake also offers solutions for the insurance and retail sectors.

Among its key partners include IBM and Microsoft with Huawei being the latest partner Silverlake is developing an AI-native core banking platform with. Each of these vendors bring unique opportunities to customers in adopting new technologies.

CRN Asia speaks to Cassandra Goh, Group CEO of Silverlake as well as Laurence Si, Managing Director, Microsoft Malaysia to understand how both the partner and vendor are providing AI to customers in the region.

How are you guys working together in providing AI to customers?

Goh: For us, we did a big reorganization within ourselves. So historically, Silverlake has been a core banking solution provider. But we think going forward, if you really believe in the capabilities of AI in the next five years, then there's no point having a single core banking application. In many senses, you break applications down to modules and you can package many things together.

Today's applications are a function of how banks' business grew over time. So, back then you bought core banking, then you bought credit card, and then you bought ATM, and you bought it all in chunks. To make things better going forward, it has to be standardized or at least have a platform across.

Because there's no point to “meng-AI-kan” your bank if you're only going to do it on a silo-by-silo basis. Then you can't get the returns on your investments. So, on the application side, we are re-jigging the way we design the product and how we want to package it together.

But that is only the top part. Underneath that is how our application talks to the boxes. How do we use the network, the tools, the different tool sets that come from the infrastructure provided.

We want to make sure that we fine-tune the application infrastructure. So again, it's not many silo pieces again, that it goes hand-in-hand. It needs to be scalable over time to take in newer ecosystems or applications that sit on top of the infra.

When we work together with Microsoft, we want to make sure the layer between the application and the infrastructure is cohesive. It is not you do your thing; I do my thing.

In the future when we grow together, we can bring in even third-party applications on top of that base. That layer exposes all of the banks to the potential of AI. So, you can really make the whole bank more productive instead of silo things.

Si: I would say that the FSI domain expertise is where Silverlake brings a lot of strength and insight. They have 35 years in the domain knowledge, and no one is going to replace that.

We provide that horizontal AI platform that they complement their domain knowledge and really come out with something that is unique and different shape in the market as well.

So, there's a lot of back and forth in terms of understanding the addition as well. It's a highly complex environment, so you can imagine there's a lot of ongoing conversation around that as well.

But then when we come back and then go back to our engineering team, which will be the right AI solution that potentially fits into what Silverlight wants to go. We potentially have to co-create and look at some new solution together as well. So, there's a lot of this ongoing.

If you look at AI, the possibility is infinite. But we should find the right use case and the right set of data to work on as well. We don't own those customer data. Silverlake has access to those data and I think that's where the synergy comes from as well. Silverlake brings a lot of out-of-the-box thinking as well into this whole thing.

Even though they're being focused on legacy banking, at the same time, I think there's a lot of out-of-the-box thinking about what the market is. They understand the complexity, but they also try to see how we unlock or undo the complexity by playing with AI to provide new insight, new experience, new hyper-personalization and more.

Do regulatory requirements have a role to play in how banks decide whether they use the cloud or remain on-prem? How complicated is that conversation?

Goh: I would say that because there are so many players in that conversation, it is quite complicated. Yes, there are regulators, and yes, there is your enterprise architect's preference, there is your CEO's preference, and then there is your boss's preference. There is also your business user's preference. That's partly the reason why Silverlake is making sure that we work closely with our infrastructure partners like Microsoft.

You have too many options already, and if the application vendor and the infrastructure vendor cannot get on the same page, the project is doomed to start.

For our joint customer together, when it comes to making the case to Bank Negara regulators, we both go in together. When it comes to making a case to the EA, we both go together. I'll be honest, a project can really go off the rails if each side says different things.

Synergy needs to be there. There is really an infinite universe of decisions that can be made, and if two sides don't at least align upfront, then the customer will help us make an infinite level of decisions.

Si: I've been working in cloud space for a long time. I think over the last decade, we have been very engaged with all the legibility. I would say Bank Negara is very progressive in terms of how they think about cloud adoptions, with all the necessary guidelines as well.

Every FSI customer tends to interpret the same guidelines very differently as well. I think that comes from different environments. Some of us come from an environment of, what is my risk exposure? Some are coming more from an angle of cost-saving or agility as well.

It really depends on what drives it as well. But we do think that there has to be continuous education with the banks. For example, a kind of explaining how this outsourcing and RMIT coexist in the area of cloud and AI as well. As the level of understanding differs by customer as well, constant engagement and education together with the vendors is important as well, especially when it comes to their first workload or their first project.

I think hanging it all together and working through that whole legitimacy and compliance process is very important. If you just say, this is a license, this is my platform, and then you walk away. I think there is no injustice to that level of engagement as well.

This is where our value-add is. We have a good relationship with our partners, we have a good relationship with the policy makers, and again with the customers, how we can jointly go to them and roll out the project together.

What is the most complicated part of conversation when it comes to AI adoption?

Goh: I think the issue is that it's not a complicated conversation, but it is very amorphous. If I say the word AI to the chairman of the board, he thinks of a sheep. And then if I say to the CEO, he's thinking of a cow. Then you say to the IT guy, he's thinking of a dog. They're not talking about the same thing. And you need them to agree on one thing to get it going.

So generally, the conversations we try to have with our customers and also for ourselves are those in the last mile of AI. For example, when you think about a chatbot or about fraud detection, let that be a third-party integration.

But the more important thing is to have your entire operations, your business and your enterprise structured in a way whereby you can host AI. This makes it easier to focus the conversation.

What about ROI? Is that part of the conversation as well?

Goh: Of course. There's a lot of ROI when we do an AI project. Oftentimes, what I try to explain to customers is there's two parts. Take chatbots for example. There is is the short-term ROI on your chatbot project and there's also a longer-term ROI on making sure that you can make your entire enterprise more efficient.

For example, Microsoft has AI agents. But if you just deploy the AI agents on segregated piecemeal projects, you won't get the full impact.

So how do you get the full impact of an enterprise-wide thing? The way that Silverlake and Microsoft have come together is to say, let's focus on one segment of your problem. If you have three, it's hard to come in from a change of infrastructure. You must solve a business problem first.

For example, if you have three existing credit card systems, consolidate it onto one. That's already one consolidated layer for you to use AI. It impacts both your IT and your business.

And since you're already there, and you've put in an enterprise platform already, consolidate again another group of common services. And then by the end of that, if you have the long-term planning for it, you have transformed your entire organization.