Synopsis: In this AI Leadership Insights, Mike Vizard talks to Sajid Husain, head of technology and innovation for SLK Software, about how AI will be used to streamline compliance processes.
Mike Vizard: Hello and welcome to the latest edition of the Techstrong AI video series. I’m your host, Mike Vizard. Today we’re with Sajid Husain, who’s head of technology and innovation for SLK Software, and we’re going to be talking about how generative AI will be applied to compliance processes that I’m sure nobody enjoys. So maybe we can figure out how to automate. Sajid, welcome to the show.
Sajid Husain: Thank you, Mike. Happy to be here.
Mike Vizard: I think there’s a lot of work that goes into compliance. It’s necessary, but it’s certainly not a lot of fun for folks and it involves massive amounts of data and processes and comparing this to that. So how do you think generative AI will change the way we think about this whole thing?
Sajid Husain: Yeah, as you said, compliance is not a fun work, so to say. And from that perspective, businesses have to do what they have to do. So having said that, in the recent context of the particularly generic coming into the picture, I believe it has started already showing a lot of result in terms of how they can ease some of the operational aspects of the regulatory compliance, whether it is on the governance side, on the risk management side or the overall compliance side itself. So technologically speaking, I would say that previously what used to be not just the mundane tasks, but also complicated tasks in the sense like how do you keep track of some of the regulatory compliance updation that happened. It could be at a geography, different countries level, or it could be even within the same location itself. How do you keep track of that and how do you keep yourself updated and how the company basically follows those compliance norms.
That use to be an uphill task. So, us companies need to invest a lot of resources both in terms of people and technology, infrastructure, et cetera, to be compliant. But now in the context of I would say gen AI, some of the difficult tasks, for example, data collection itself, how do you ensure that you keep track of some of the regulatory compliances itself? That has become easy, right? That’s one part. Second is how do you interpret some of the new laws or the regulations that have come up? So again, since a lot of it is related to legal jargon, et cetera. So gen AI can play a part in even making it easy for in other companies to understand what kind of new regulation that has been updated, how is it different from the previous regulation that were there, et cetera, right? So that’s the second part.
Third part is I would say is specifically more in the context of the data collection and interpretation itself. So not just the data which are part of the company themselves, but also the external data that is from an industry point of view also, particularly if you’re looking at GDPR, HIPPA and so on and so forth. So again, gen AI can help in collating the data, but also interpreting the data, which again makes it very easy because again, if you’re talking about hundreds of thousands of data sets, we are talking about hundreds of documents right now, going through them, understanding them, interpreting them, and then bringing it to, from a compliance point of view for the companies, it’s a major task and then a challenging task.
So that, again, generative AI can ease some of the aspects of by virtue of, again, NLP is one of the strength of gen AI. So it can scan through all those documents, try to interpret the compliance related to that particular industry. It can even compare to some of the compliances that need to be adhered to for each of those companies. And it can automatically provide non-compliance indications much before it really happens.
Mike Vizard: Do you think it’ll become simpler for mere mortals to understand what’s in these compliance documents? Because I feel like they’re generally created by lawyers who appear to be getting paid by the word, and these things are incredibly long and maybe an easy summarization would just make it simpler for everybody to wrap their head around what needs to be done.
Sajid Husain: Absolutely. I think, again, as I mentioned in the context of NLP, that’s one of the strength of gen AI. So in fact, not only consuming the compliance related documents and then trying to interpret, some of them, in fact, if you see within the US for example, some of the law firms have already started providing the services to their customers using gene AI as in creating even the contract documents and then the compliance related documents, et cetera can also be generated as well. So it can be consumed on both sides, one on creation for it in terms of wherever it is required. The other is consuming it and then trying to interpret it and then trying to make it easy and simple and understandable format.
And also not only just making it summarization, but also giving pointers, saying that, okay, by virtue of comparing the data within the company, it can even start giving pointers saying that, okay, for example, some of the compliance has saw the threshold of almost breaching, which could result in huge fines. So gen AI today can automatically start looking into the data from a compliance perspective and then try to warn the companies saying that, okay, these are breaching another kind of level that is almost not permissible. So not only it can interpret, but also it can help proactively even to identify and then trying to prevent any non-compliances.
Mike Vizard: Do you think we’re going to get to the point where gen AI will help us to automatically create the control that’s required? Not just tell me the summarization, but a lot of people are talking about compliance as code these days, and I really just need a developer to create something that I can insert into a workflow. I mean, are we on the cusp of that?
Sajid Husain: In fact, it’s already happening. I mean, today, as I mentioned, I know in the beginning… So the data collection that is happening from the external sources, that itself is a good input for the companies to compare, and particularly for global companies, let’s say if you take supply chain as an example, right? Or healthcare companies where they have to adhere to the regional aspects of the compliances across the globe. So they already are utilizing it to consume the data and also pinpoint it and also provide them some kind of mechanism for adhering it as we go forward. And also it even helps even the companies, the employees, even to train them also.
I mean, particularly when any compliance related changes happen, gen AI can detect that and it can even publish it across the organization to talk about, particularly with the respective teams, whether it’s legal or the compliance team. To proactively say that, okay, these are all the new change that will happen and these are all the control measures that you need to bring in and some of these levers that you need to control. Let’s say if it is related to supply chain and climate change activities, so there, it can literally track that [inaudible 00:07:05] carbon emission, et cetera, for example, and then compare it with the compliance related things. And it can proactively point out saying that this is something that needs to be looked into a little bit more in detail compared to what it was let’s say a few quarters before, because the compliance regulations have changed.
Mike Vizard: Do you think this will change the politics of compliance? And I’m asking the question because the debate has always been we need more regulations on one side, and then the business is saying it’s too costly to implement and it will drive everybody out of business and we’ll never get anything done. So can we get to the point now where we can intelligently apply regulations without necessarily driving cost up through the roof?
Sajid Husain: In fact, it’s the other way I would say. I mean, it can help a lot in reducing the cost for the companies to be compliant, which means if you had to invest a lot, particularly for small and medium-sized companies, if you have to be really compliant, particularly let’s say in healthcare and financial services companies, previously, they would’ve had to invest a lot in terms of infrastructure, people and then the team they have to build in, whether through legal or the compliance team, et cetera. Now that can be hugely reduced by automating some of the tasks which used to be done either manually or semi-automatically, semi-automous way. That can be today automated completely with the help of gen AI. And more importantly, again, as I mentioned, it can not only complete those tasks automatically, but start interpreting those also to help augment the business processes itself for the companies to make it easy.
So I would definitely say that the cost will, again, I can’t put a number to it, but I would say that definitely by virtue of automating a lot of these processes and linking it to the existing business process of compliance, it definitely will drive the cost down. So it’s definitely going to be an advantage for the companies to utilize gen AI to be compliant.
Mike Vizard: Do you think that as we go along here, there’s a best practice for getting started? Because one of the things that I feel, at least when I talk to people is they’re overwhelmed by compliance as is, but they’re also overwhelmed by AI and they don’t know where to get started and they don’t know how to get this journey going. So what’s your best advice to folks?
Sajid Husain: Well, okay, that’s a good question and I can start by saying that [inaudible 00:09:30] can help you, but the fact is that if you look at from an adoption point of view, companies definitely can take different types of approaches, and particularly for… It all depends on two things I would say. One is the understanding of AI engineer within the company itself, which basically means you’re talking about skills in terms of the capabilities. The second part is the cost. So looking at both as long as the companies can put up a framework, for example, again, if I can just quote an example of how SLK helps our customers to help in this journey, typically we have developed what we call as an enterprise adoption framework. Now, interface adoption framework basically talks about multiple phases, how you can go about starting from ideation to execution to sustenance. Now, in each of those phases, literally you have to go through some of the processes.
You have to understand what needs to be automated, where AI can make a difference, and what will be the ROI for each one of these. Because at the end of the day, gen AI also comes with its own technological overhead with respect to cost. So if you bring in gen AI will that have an impact with respect to cost reduction for the company. Those considerations has to be done. Now based on that… And of course, one more, I would say most critical thing is to have all the stakeholders within the company participate in the strategy for how and when the adoption has to happen. So if you look at, for example, and if you look at manufacturing industry, they have been one of the first to adopt because for them it makes a huge difference in terms of from an automation perspective, supply chain perspective, supply chain and logistics perspective.
Whereas if you look at some of the other industries where the regulation is very high, including banks and financial services companies, they have been very, very cautious, even though they do understand some of the benefits, huge benefits I would say that it can bring in, but purely because of the regulatory constraints, they’re trying to look at it from a wait and watch approach. But having said that, again, in the last I would say six months, we have seen quite a bit of, I would say financial services companies and even healthcare companies that are trying to adopt gen AI in a small way, which according to me is the right step to do. So you start small, have some kind of a planned execution, a plan with respect to the timeline, right? Start small, deliver results, and then move to the slightly bigger augmenting the business processes and then move on to in a much, I mean completely automation of the business processes using AI.
Mike Vizard: Can we connect the dots between the violation and the potential cost better than we have in the past? I think a lot of times people are like, oh, well, we’re going to be in violation of this, that, and the other, but they don’t really know what the potential cost is. So can that become something that AI surfaces, at the moment, we’re about to violate said compliance protocol?
Sajid Husain: I would say yes, because as long as, let’s say if the company, from a process point of view, they understand that any kind of violation with respect to the compliance, what is from a regulatory point of view, what kind of fines that they’re going to incur or what kind of stipulation that they’re going to incur from the government or the necessary companies. So they have to be prepared with that kind of data. So I would say as long as the company is aware of it in terms of how and what is the implications of not adhering from a compliance point of view, they can easily have some kind of a simulation. For example, if you look at manufacturing companies, so they call something called a digital twin. Now digital twin is a thing, but whatever the business process you have, manufacturing process, you’re trying to do the same thing in the concept from a digital point of view. So that instead of trying to build a new factory and try out, you actually try it out digitally.
Now, in the same way, the businesses can do the same thing that, okay, I can simulate it in case if there is a non-compliant that happens, but what would be the impact on the business? People have actually built those models already. It’s not new, which means that for all practical purposes, the company have to be prepared because sometimes it’s quite possible that they could be noncompliant simply because of, not necessarily because of negligence, but because it could be because of certain circumstances which are out of the control. So in that context, they need to be prepared ahead of time. And I would say that definitely there are a lot of models that are available for people to simulate it and then try out what happens in case of the company stand out to be non-compliant.
Mike Vizard: You’ve been at this a while, what’s the one thing you see people doing as it relates to compliance over and over again that just makes you shake your head and go, folks, we’re better than this?
Sajid Husain: Well, I would say that particularly, I mean, if you look at internet, if you look at even the companies, the amount of data that gets generated is huge and on a daily basis. Now that particular data, the data intelligence, both from internal sources as well as the external sources, bringing it together to look at it, not just from a compliance perspective, but also using it for your competition management or for your best business practices, et cetera. I would say that non utilization of data practically or properly is one of, I would say the weak points, weak link in the chain. Now, having said that, again, things are changing now from a data interpretation perspective, data consumption perspective. As I mentioned, gen AI has started playing a big role and now slowly we are seeing the adoption of consuming of that data, excuse me, and passing on the so-called data pipeline to gen AI, and then using the applications to derive insights out of it, which would not have been possible before, I would say is one of the biggest challenge, which again, things are moving in the right direction.
Mike Vizard: All right, folks, you heard it here. I wouldn’t say that AI is about to make compliance joyful, but it’s certainly going to be a lot better than it’s ever been in the past, and maybe we might actually learn something as opposed to just checking a bunch of boxes. Hey, Sajid, thanks for being on the show.
Sajid Husain: Thank you, Mike. Thanks for inviting me here.
Mike Vizard: All right. Thank you all for watching the latest episode of the Techstrong.ai Video Series. You can find this episode and others on our website. We invite you to check them all out. Till then, we’ll see you next time.