TechCentral (main feed)

TCS+ | Asokan Moodley on Nedbank’s journey into AI

Asokan Moodley is leading Nedbank’s drive into the world of AI – and it is already starting to have a meaningful impact on productivity in the bank. In this episode of TechCentral’s TCS+, Moodley – Nedbank Group Technology’s head of end user and communication experience, infrastructure and operations – unpacks what the bank is doing with Copilot for Microsoft 365 and the lessons learnt so far from the deployment of the technology. Nedbank, which is one of the first organisations in South Africa to deploy Copilot for Microsoft 365, has made the technology available to select employees, including senior management, ahead of a planned wider deployment. In the interview, Moodley discusses: • Why Nedbank decided to be a relatively early adopter of Copilot for Microsoft 365; • Which employees the technology has been deployed to initially and why they were chosen; • The business challenges the bank is hoping the technology will address; • The important lessons learnt so far, and how other companies should be approaching their deployments; • The potential pitfalls of using advanced AI tools in a highly regulated industry like banking; • The security concerns, and how Nedbank is addressing these; and • How the bank helped employees embrace AI tools in their day-to-day work, and the feedback they have provided regarding their use of the technology. Moodley also shares his views on whether AI tools, including Copilot for Microsoft 365, provide a competitive edge in both the short and the long term. If your organisation is thinking of deploying AI technology to its employees – and especially if it’s considering Copilot for Microsoft 365 – this is a conversation you shouldn’t miss, especially as Moodley shares his views on how leaders in other organisations who have embarked on a similar journey should approach this.

Duration:: 26m
Broadcast on:: 28 Jun 2024
Audio Format:: aac

[MUSIC] >> I'm Duncan McLeod, and this is Tech Central's TCS Plus, the Business Technology Show, which is brought to you today by Microsoft and Ned Bank. Now, I'm actually very excited about this next conversation, because we're going to be doing a bit of a deep dive, if you like, into Ned Bank's use of Microsoft 365 co-pilot, and it's one of the first companies to deploy the artificial intelligence technology broadly in its operations here in South Africa. So I think there's going to be an enormous interest in this discussion from other companies as well, who are looking at deploying co-pilot, and I'm wondering what it's all about, and what's involved, et cetera, et cetera. So to do this, I'm joined from the UK today. Actually, he's in Ireland, in Dublin, I believe, just returned from the UK on holiday. Thanks for interrupting your holiday for us, Asakan Woodley, who is executive head of end user and communication experience, infrastructure and operations in Ned Bank Group Technology. It's good to see you today. How's Dublin? Good morning. Thank you for the opportunity. Dublin's fantastic. Wheeler is great today, so, yeah. Sunshine, no cloud, no rain, which is good. Excellent. Well, Asakan, I'm looking forward to the discussion, and we won't keep you too long, because I'm sure you want to get back to your holiday. But thanks for interrupting your vacation for us, and chatting a little bit about AI, and how Ned Bank is using the technology. It's been about 18 months now, since Open AI broke cover, with chat GPT 3.5, and of course, Microsoft has been working feverishly since then, to introduce a generative AI tools across its operating system, and its business software now. Ned Bank has decided it's going to be a relatively earlier adoption of the technology here in South Africa. Maybe take us through what led to Ned Bank's decision to deploy co-pilots for Microsoft 365 in the bank, who's using it, and what they're using it for. Yeah, thanks, Sun Ned Bank, when decided to go into the early adoption program with Microsoft in terms of its use of co-pilot, we procured about 300 license initially, just to try it out, to see what improvements, what innovation we're going to get across the bank by using co-pilot. One or two of the challenges that we were facing in the bank was, and an example would be meetings, secretariat, sitting meetings with recording a meeting, and they've got to go back to their desk, and listen to the recording to draw up minutes. So what we've done is, from a usage perspective, we kind of took it across our business units, and Ned Bank has various business units across South Africa, UK, and into Africa and Isle of Man. So each of the business units across Ned Bank were selected, where we had users. We call them champion users across the organization, identified by each of the CIOs, and they were then enabled with these licenses to be able to use it in a way that we were trying to improve productivity across the bank. And what we've seen is that these users have benefited tremendously from the licenses that they have thus far. So in terms of where we deployed, it is from management to back office users, to the secretariat in the organization. And where have you seen the biggest productivity impact so far? I think there's mainly two areas that we've seen it across the organization being on developer community. Program is right cold, and with co-pilot, or GitHub co-pilot as well. It's the other product that we deployed. Sorry. We've seen a tremendous gain in productivity in a developer being able to write code a lot quicker, because now you're using the co-pilot to do it for you versus you having to go through tremendous amount of coding per line. It reduced their time to delivery by almost 70% in the organization. Other areas from a M.S. co-pilot specifically, the secretaries across the organization are now using co-pilot for doing their minute taking. So they turn on transcript during the meeting, because co-pilot allows you to do that. And at the end of the meeting, they do notes in terms of what the meeting content was about, and obviously it being co-pilot, and they call it co-pilot for a specific reason. Everything's not correct. When you do the transcript, there might be some name incorrect, but it kind of gives you a good idea of what the meeting was about, and the secretaries take that straight into the document, fix it a little bit, and where they've taken a day or two to do minutes, it's now taking them 30 minutes to do minutes. I mean, other areas of improvement we're seeing is in our investment banking environment, where from a client perspective, drawing up specific documents for clients, they're now using co-pilot to do that. So if they have any client contract, as an example, co-pilot summarizes it for them, and gives them the main essence of what the contract is, that they can then engage customers. There's various other use cases. I could go on all day talking about use cases that we have in the bank. But what we found is there are common use cases across all business units. We picked up at least four common business cases across, it doesn't matter if we start expanding the co-pilot across the organization from a licensed perspective. We know those four standard use cases are going to give us immediate benefit across the bank. What about at a senior management level, including in the C-suite? Have you deployed co-pilots to those people, and are you seeing the music in interesting ways as well? Interesting question. So we have deployed to some of the C-suite community, like our CIO and our board members, some of them have it, and now that Ned Bank has a new chairman and a CEO, Jason McRin, we've deployed it to them as well. Some of them use it to the extent that we would like them to use it, and they are definitely seeing benefit. For them, especially from a C-suite perspective, they try and maximize their time. What we're finding now for them to do a presentation is an example. They have content sitting in various documents, word documents, using co-pilot reference those documents and ask it to do a summary presentation for them, and co-pilot builds a presentation within two minutes from content that could be a 30-page long content, as an example. All the other areas they're using it is if they were late or somebody was late to a meeting, you join the meeting, and you don't know what happened in the last five minutes of the meeting because you came in late. You can ask co-pilot to give you a summary of what happened in the last five minutes prior to joining. It'll tell you exactly what happened within the action items for you from that meeting. There's massive benefit from that. It's very innovative and intuitive. Do you see it being used at some point for making decisions as well in the organization, at a senior level, ingesting a bunch of data and saying to co-pilot, what is the optimal solution to this, or can you help me devise a strategy or a decision, or come to a decision around a particular issue, or do you think the technology is not quite that advanced yet? I think making decisions using co-pilot, I don't think it says as advanced to that level yet, but what it can help you do, though, is give you enough content to make the right decision. It could, as an example, from a financial perspective, you could ask co-pilot, from an organizational and specifically NEDBEN. Tell me what our ROI and CPI was for the last three years across the organization, and it will go pick that content for you and give you information from whatever, you know, 2022, 2023, and if you're doing your financial year, it'll go find that information and give you a summary, so it'll help you in meaningful content like that to make the right decision. Okay. What are the most important lessons that NEDBEN has learned from deploying this technology so far, or is it too early to really draw big lessons at the stage? I think there are some positives and some negatives from a learning perspective. The positive side of it is that it is certainly we see an improvement in productivity in quite a few areas. What we've seen is that it's brought on unparalleled innovation and efficiencies together with data protection in the organization, so definitely seeing that self-productivity in terms of the 300 that have the licenses as improved. Now, banks are obviously highly regulated entities. How did that factor into the way that NEDBEN has decided to deploy co-pilot? I think, from that process perspective, before we could even enable these 300 users, we had to go through a harder, rigorous governance process to make sure that we have certain guardrails, policies in place before enabling. AI specifically is relatively new in the market. There's not a lot from a South African perspective in terms of regulatory bodies, but we took the stance of following the international EU regulatory bodies and standards and took some of the lessons out of what they've done from a data governance perspective. Does it's key from a banking or any other organization for that matter when you start using AI, that your data is still controlled within the organization and not put externally out there for other people to use? We had to go through a process from a security locking down certain instances. Because we started adopting AI in the organization, the co-pilot built into the Microsoft Web Engine, we decided as then to lock all other light set GPT as an example, lock it down or lock and not allow users to use that because you could put your data out there. So, the way we've coined it is we able to use the internet, but the internet is not able to use us because our data stays with us. We've gone through that process. We've created user acceptable guideline policies that we've now sent out through the organization so that users know what they can and can't do with AI, what information they can share and what they can't share. So, from a governance perspective, I think it was quite a process that we had to go through and educating people on what they can do. So, do you think there's more need in South Africa or a need in South Africa for more regulations to maybe to align us more with what the EU is doing? Yes, certainly. Look, there is a regulatory body now formed in South Africa and specifically, I think the organization sits in Cape Town, but I think there has to be a more active role from key participants across the country that should participate in those forums so that the right governance and regulatory can be drawn up and circulated from a South African perspective. So, there is an organization that I think we're just a little bit slower than the EU to take off. What has been the feedback from the users you've deployed the technology to so far? I imagine that some employees are a little resistant and just natural human resistance to change while others say this is fantastic. I can see others are going to change my productivity and jump in head first. Have you ever seen that sort of dichotomy within the organization? Have you had to help some of the users along the process as well with some training and helping them understand how the technology can help them? I think that's actually a very good question there, because with any AI that you bring into any environment, the first thing people ask you is my job underlying. Am I going to lose my job because AI is now going to do my work? It's not so. We have to go through a process of adoption and change management, which I think helps. So, if people are going to be deploying co-pilot in the organization, make sure you follow the adoption and change management process across so that you're educating people on how to use the co-pilot and what you're going to get out of it. Because giving somebody a tool and not knowing how to use it, you may interpret it incorrectly or get the incorrect data out of it, which could be a problem. So, we went through a process of what we call, we were holding bite-sized learning sessions on a weekly basis with the users so that they could learn. We could teach them what co-pilot does, what some of the prompting that you can do. I think that's important, prompting in co-pilot is very important because if you ask it the right question, it will give you a meaningful response. You can train it eventually in terms of how you as an individual ask questions. It will learn your style of asking questions, but what it doesn't do is store that information in the club for external users to gain access to it. So, that training was important and getting people's mindset to change that your jobs are not on the line, but co-pilot is here to aid you and improve your productivity and innovation across the organization, which then it makes the individual's job easier as well as again for the company because you're now getting more out of the individual. But feedback from individual users, what are they actually telling you about co-pilots? Are they excited about using it? Are you seeing having tangible benefits? I mean, you spoke about the transcriptions for the secretaries. I mean, you've got guys coming back to you within the organization saying, "I don't know how I worked with others' technology previously." Yeah, we've actually had users threaten us to say, "If you take this license away, I'm going to go start toy-tuning and placarding outside the organization." But as a joke, they said it, right? But certainly, we've seen a huge positive feedback from users across the organization that have the license in terms of how it's really changed the way they work. And yeah, a lot of them are saying they're not sure how they work previously without us because having to go through a 30-page document, and at the end of it trying to summarize it, now you do eat the document because you want to know what it's counting. But you ask co-pilot to give you the key salient points out of it. It just makes life easier. You ask co-pilot, "Tell me, were there any actions for me in the meeting at all? I immediately tell you." So the users are saying, "Well, what a fantastic product." And it's certainly making their lives easier. So a lot of positive feedback. Excellent. Do you think that in the longer-term deploying AI and deploying it early like NetBank is doing will give you a competitive edge? Or do you think this is just the ticket to the game now that everyone's going to deploy co-violet or some equivalent technology? And it's just the entry ticket to doing business in future. So I think there's twofold. What AI, generative AI, co-pilot being relatively new in the market, if you're adopted early, it certainly gives you a competitive advantage in the market. But I think going forward, it's going to become a norm. Every organization will probably start adopting it, and it becomes part of your day-to-day lives going forward. So I think if companies are given the opportunity, they should start using co-pilot because it certainly gives you some competitive advantage. Okay. Now, how does deploying AI compared to a digital transformation project? Many South African organizations have been through digital transformation projects. I know that's a wide-ranging term. Do you believe they require similar frameworks to be successful? Definitely. I think if you look at really digital transformation, and it depends how you see digital transformation, digital transformation could mean that you are rewriting applications to be more user-friendly on devices like mobile devices, iPads. Whereas previously, you would have had to go from a banking perspective going to a branch to do certain things. Now you can transact and do things on your mobile devices. There's no need to go into a branch. So for me, you know, digital transformation, from an AI perspective, I think it's going to assist in people that are still lagging on digital transformation because AI is not going to help them speed up their process and help with modernization a lot faster. Okay. What advice would you give to other companies who are listening to this, who are thinking of going down this journey themselves? Can you provide maybe some pointers to leaders of other companies, other industries, on how they can ensure they meet the expectations of the employees around the deployment of AI in the workplace, from the lessons you've learned as net banks so far? So I think the key components here from what other organizations should look at doing is one would be make sure you have the right governance in place when you're adopting AI and just don't give it out there freely hoping that the user knows what to do with it and how the data is protected. So from a governance perspective, make sure you tick that box off first and put in the right data protection policies or go relook at your organizational policies and structures that you have in place. What we find is that you have your chief information security officer as an example that performs certain tasks now with AI. That's something new that you'll see so in the organizations also taken the longest journey with you because they change and your risk and governance people are on board in terms of what needs to be done. So I think that's probably your first thing. The second would be adoption and change management across the organization is key. Make sure that that process is followed throughout the journey and not just do it once. It's like learning how to drive a car. You have to have a few lessons before you can actually go get your license for a car. So it's similar thing. Take your users along the journey and they're comfortable with using the product. Run your, what do you call it, prompt engineering courses. Make if there are any internet based training, make that available for your users. So, yeah, those would be some of the key takeouts. And lastly, so what's next in terms of Ned Bank's AI journey? Where do you see this technology taking the bank in the next few years? I think it's going to help us transform our interaction with our customers. It's going to make our front office or customer facing retail staff and our investment banking staff, our wealth staff that engage on a day-to-day basis with their clients. It's going to make their lives a lot easier from being able to deliver a lot faster to our in-custom. And I think that's the important, you know, we in Ned Bank, and I'm sure other industry talks about customer experience, digital experience, and employee experience. So we call it the CX, GX, and EX. So that's certainly going to start improving those environments or those areas specifically for our end users, both external and internal. Fantastic, very interesting discussion. Look forward to seeing the impact on Ned Bank and its customers going forward. Ahsokan Woodley is executive head of end user and communication experience, infrastructure, and operations in Ned Bank Group technology. Thanks so much for joining us all the way from Dublin today, and do enjoy the rest of your holiday. Thank you very much. Great chatting with you, and hope people find this useful. [BLANK_AUDIO]