In this episode of Leaders of Code, host Ben Popper chats with Prashanth Chandrasekar, CEO of Stack Overflow, and guests from Lloyds Banking Group, Hilary Lanham, Head of Engineering Excellence, and Tom Kelk, Senior Product Owner, Productivity Intelligence, about their focus on engineering excellence and the need for organizations—particularly those in highly regulated environments—to adapt to new technologies while ensuring customer safety and meeting their expectations.
They also discuss:
Episode notes:
[intro music plays]
Ben Popper Hello, everybody. Welcome back to the Stack Overflow Podcast, a place to talk all things software and technology. I am your host, Ben Popper, coming to you with a slightly baritone edition. Got a bit of a head cold, but that's all right. Today is a Leaders of Code episode. I'm going to be hanging out with Prashanth Chandrasekar, CEO of Stack Overflow, and some folks from Lloyds Bank who we have done a lot of work with in terms of Stack Overflow for Teams and who have a lot to tell us about engineering excellence and what it's like to work in a highly regulated industry. So without further ado, Prashanth, welcome back to the show.
Prashanth Chandrasekar Thank you, Ben. Wonderful to always be here. And welcome to Tom and Hilary. Great to see you both after seeing you both in person about, I think like a year and a half ago in London.
BP So Hills, why don't we start with you. Tell the audience who you are, how you found your way into the world of technology and finance, and what it is you do day to day in your current role.
Hilary Lanham My name is Hilary Lanham. I live in Bristol. I joined Lloyds Banking Group 20 years ago now. I got into Lloyds through being effectively a temp employee straight from uni and here I am 20 years later. It's hard for me to believe it's been that long. My current role in the bank is Head of Engineering Experience in a wider engineering platform we've got, and our products are focused on making the engineers’ lives easier in Lloyds Banking Group. It's less around the technologies they need to do their jobs, more to do with the products to how to understand how they feel about doing their jobs in Lloyds and what we can do to make their lives easier.
BP Developer happiness, key metric if you want to keep things running. Tom, how about yourself?
Tom Kelk I had a bit of a strange route into the bank, so I've been at the bank 10 years. I used to work in sports marketing. Digital marketing was always my thing. I joined Lloyds and realized that subjectivity of marketing wasn't something I could hack. Went more towards technology and marketing, technology operations, and then more and more into the organization and the technology operations of it. Worked into the chief technology officer for a little while and then found my way out to product ownership ultimately, which is where I am at the moment, working in the engineering experience team that Hills outlined.
BP Very cool. I'm not sure that I quite know what that title means. What is product? How do you define product ownership?
TK Very good question and very pertinent to the moment as it's all being redrawn. I was actually at a product owner meetup in Cambridge last night and it is interesting that everyone has a different definition of what a product owner or product management looks like, whether product is a journey end to end of someone joining an organization to leaving it, or whether it's something someone consumes. So in the organization Stack Overflow, we brought Stack Overflow in as a product, and I think the difference people generally look at with project to product is project is the delivery of something, the product is almost the fun stuff afterwards. How do you actually get someone to use it and how do you get people to enhance it, and that's the experience we've had within the group. So I see myself as the latter part to delivery, the afterburn if you like, where things get, in my mind, a little bit more fun.
BP In my experience over the last 10 years, often you find yourself at a company, you're new there, you're getting onboarded and there's 85 different SaaS tools that they have that do different things, some of which seem like they're overlapping and some of which seem maybe not to have been updated in a long time. And so when I think of product owner also, it's who's making sure that these tools we're using are actually getting used, that the experience is good, and that folks who are then going to come in and touch them are not going to be using something which may be outdated or impossible to work with.
PC Yeah. I mean, this is a multi-decade problem that you're talking about, right? It's collecting dust on dust which I think needs to be cleaned out and spring cleaning should always happen along the way. So yes, I recognize the problem for sure.
BP What are some of the things that you see as areas of friction or managing adoption or trying to integrate some of what's cutting edge these days, but making sure you do it in a way that's respectful of data and privacy, especially within an industry like yours?
HL So the challenge you just outlined is absolutely what we are dealing with at the moment– making the onboarding journey as smooth as possible. But as people join our bank, their expectations are very different. We might be a highly regulated centuries old bank, but when engineers join us, their expectations are cutting edge technology, fast access to data. They want to be in charge of getting what they need with a zero ticket process, very self-serve. And of course, in a highly regulated environment, that's difficult. We've got security and regulators to contend with as well. So I think one of the big challenges is managing the balance between engineers' expectations and what we can offer them in terms of the freedom of access that they perhaps would be able to experience in different companies to what we've got in Lloyds.
TK Yeah, I think what you were highlighting earlier about the hundreds of different SaaS solutions available to our engineers is an experience we've had historically and we continue to have. And where Hills talks about improving the experience of the end user and making sure that onboarding is seamless and quick, you can't do that when you have so many solutions which need integrating together. So the biggest challenge we are facing at the moment is simplification, and that's the biggest strategic imperative for the organization right now. How do we simplify down the tools that we use so that, back to the product owner conversation, we can enhance the experience of one or two things as opposed to try and enhance the experience of hundreds.
BP I know you're working on a project at the moment that involves Stack Overflow for Teams. Maybe tell us a little bit about that, what you're excited about, and then you and Prashanth can talk a little bit about some of the ways in which new technologies like LLMs and various aspects of generative AI are allowing us to consume data and knowledge new ways, and sort of increase the circulation of some of that stuff inside of an organization in a way that can be really productive.
TK Of course. So about 18 months ago/two years brought Stack Overflow for Teams into Lloyds, and I think at the time the understanding of what it could provide was simply that we needed to break down all of the many different silos we had in the organization. We had 30+ different platforms of people which contained thousands, and often I think it was 60% of those people hadn't spoken to anyone outside of their team since joining. So we needed to bring in something that could help break that barrier down, but also try and bring the knowledge out from all of those individuals and enhance the overall knowledge within the organization. So we've succeeded in doing all of those good things that Stack Overflow for Teams is there to do. I think what we're really excited about going forward is the data within Stack Overflow for Teams, but also the data that we hold as an organization and how we can use that to increase productivity and remove blockers. And I spoke about product ownership, but what Stack Overflow does really well is it brings all of the conversations that are being had about individual products across the organization into one place, and that allows us the opportunity to take that data and then curate it for product owners wherever they are. And so a product owner doesn't now have to go and speak to a few louder people who they normally speak to to find out how the experience of their product is. We can give them an unfiltered view of what people are searching for about their product, what people are saying and asking questions of, and who is answering the questions that should be answered by those product owners, by the SMEs. We can give them a rich picture that we've never been able to do before and we can supplement that data with location based specific agenda stuff. We can do that with our survey data and pull pieces of kind of a rich persona at the individuals at the end of our products together. So I'm really excited about where this goes going forward because productivity, measurement of productivity, but then also the data that supplements all of those conversations is the most important thing in my world this year. So I'm really excited about where it can go.
PC Yeah, amazing. Tom, I think the point you're making with the fact that you're able to use all this data in a productive way in this go forward world, it resonates a lot, only because when we think about the problems we're solving, it's about, number one, for users and our customers to be able to find the right information, and when they do find the right information, hoping that it's actually accurate and up to date. So even if you accomplish one of those things, the other thing is a limiter to be able to succeed. So the great news about I think where we're all headed is if you have a high quality knowledge base within your company or data is high quality and structured and so on, then you can rely on it and its accuracy is high because it can be attributed, et cetera, to a kind of an authoritative source within the company and it's up to date, then that is really, really valuable as you use it for all of the use cases you described. And especially as folks think about using AI agents and assistants, et cetera, I think that becomes supremely useful. So a lot of our customers, like you, are embarking on this. How do they leverage high quality data within the company, and if they haven't done it, then let's get back to basics. Let's establish a high quality data foundation and obviously our solution with Stack Overflow for Teams enables that as a knowledge store, if you will. That in combination with really powerful data sets like even our public dataset which is Stack Overflow, is now in demand by customers to be able to leverage that through an API to then augment their internal Stack Overflow for Teams and then deliver all that information at the right place, right time for users within the company, along with all sorts of other data that, as you mentioned you could augment and really personalize it based on the user. So I think it's just to your point a very exciting time because I think the possibilities have expanded dramatically all of a sudden because it's just a lot more accessible than before.
BP I love the idea, kind of to your point, of people getting this unfiltered look at what people are asking or complaining or complimenting about their tools and products through the questions that are being asked to the system, and then that can come back to them and they can take that and understand, “Okay, well people in this department are really struggling with this integration. I've noticed an uptick here, or people are really struggling since we rolled back X,” and that may be saving a ton of internal communication, but maybe even communication people don't feel comfortable having honestly if they don't know each other well or they're in different branches of the business.
HL It absolutely closes that feedback loop because in order for us to focus on our engineering experience, we need to treat our engineers as our customers. And so how else are we going to get that really quick feedback loop? And quite often we will cascade information and we won't really know if it lands or not, whereas if we feel we've delivered a message, for example, a how-to guide or best practice documentation, and yet we're still seeing a lot of questions coming in on that subject, that lets us know that we've not hit the mark from our customer feedback there. It's really important, and as Tom says, that richness of data that we are getting from Stack and from our survey results, we are not getting it elsewhere in the bank so we are able to be really powerful making data-driven decisions and basing our new features on our products based on this data of feedback we are getting. That is what's taking us to the next level in terms of serving our customers.
PC Excellent. I was going to maybe mention or ask you actually, Hills and Tom, around the fact that you're in a bank, it's highly regulated, especially in the UK, I think it's even more so than other places. And I think, I'm just curious, you briefly touched on this, but I think it cannot be understated because it is actually a large challenge in terms of being able to leverage new technology with data in the middle of it, especially data sets, which there's obviously all sorts of concerns around that. Do you think that that's changing in the financial services sector as you folks work with other banks in the industry and as people are trying to sort of really push the envelope, do you think it's going to open up? Is there an openness, do you think, to leveraging data and tools and really what we've been talking about in a much more accelerated fashion, and are you optimistic about it?
HL Yeah, absolutely. So we are genuinely seeing a culture change with our security office and with our regulators. As cloud technology came in, as AI came in, as all these large language models, et cetera, not only have engineers got to upskill, but our security colleagues also have to upskill. They need to stay ahead of the game. They need to understand the risks. And if we said no to everything, we would be holding ourselves back as an organization. So instead of saying no, the culture is, how do we be more confident in saying yes? And so we are really seeing a big change in our security posture on this where we are trying to be as amenable to all these changes as we can be whilst also always keeping our customers’ safety at the forefront of our minds.
TK And I think what Hills is saying is so important, that we expedite these processes to a point, but then we are really being tested on this at the moment with the continuing changes to large language models and what's available. We look at stuff about architecture and the ability to make sensible decisions about what our architecture looks like, but then every single day a new large language model comes out that has beaten the previous one. And we have big strategic partners that we work with and we are really fortunate to have them, but how do you design around what you provide to your engineers, what tools they use, what training they have when every single day something brand new is coming out. And our priority has to be how do you protect your customers? We have more than 30 million customers and we're seeing so many. I think it was 5 billion logons onto our mobile app. So we are seeing that interaction and you can't play fast and loose with that, but then you also have to match those customer expectations about the services that are provided by our competitors. So we have this real challenge, and probably more so now than ever with AI and what technologies you build real structural foundations around. It's a really interesting time to be, I think, in financial services.
PC We talk a lot about the technologists and the developer experience and the happiness for those folks, especially in this context, but I'm curious in the other departments across Lloyds, given what you just said about your 30 million customers, do you sense that there's going to be a lot of folks embracing all the technology concepts we covered in other parts beyond just the developer and technologist group within the company? So for example, in marketing, it could be in risk management, in other departments that are more customer-facing, as an example.
TK Absolutely. And so as I mentioned, I'm a former marketer, so I’ve still got quite strong connections in that part of the organization. They're really big on what the opportunity is with AI to enhance what we do. And I think there's so many different maturity levels and so we are spending a lot of time training people in the organization on how to prompt, what value can you get out the back of it, and also not using it just to shortcut, because I think what you see in, I'll go back to product management again, what you see a lot is people going, “Well, what's the solution?” and using AI to fast track to the solution as opposed to actually using AI to help on the data gathering and the distilling of what the idea is. And I think that's more true in marketing as well, that you could easily fast track to marketing collateral that is sent out to your customers, but actually, is that really getting to the matter of the problem? Probably isn't. So we have to educate on how you use it, but also when and why you use it too. But the appetite is absolutely there.
BP Yeah, I've heard about this challenge within a lot of organizations. It seems like what a lot of folks are trying to do is to create both a sort of internal firewall, which is to say if people are dealing with LLMs or they're having data and responses sent back and forth, let's make sure that we understand what can go out and what's acceptable to come back in. And then there's almost like a router, a switcher on the outside that's like, “You can choose today to try this model. You can choose today to try the new. Let's keep adding the latest model and seeing which one gives us the best outcomes.” These days, sometimes they go offline, they hit their global, they release their newest version and they're down for a day. Great, let's fall back on one of these other ones. So it is a very fascinating time and hard to keep up with all the changes that are happening when everybody keeps putting out new charts showing they're the best. There's always a baseline and the latest model is always so much better than the last one, but it's only three weeks away from being not cutting edge anymore. So one question I'd love to ask is a project or priority you're focused on this year that you think will have some impact on engineering sort of efficiency. Hills, is there anything you're working on?
HL Absolutely. So as a bank, we're going through a multi-year journey, driving simplification and standardization. The focus is engineering excellence, and we believe that to achieve engineering excellence, we need to get our engineers working more smoothly, reducing the friction, reducing the cognitive load for our engineers. And we are doing that through standardization. So to give you an example, we've got multiple GitHub organizations in the bank. We've got 80 federated artifact repositories. We really need to start reducing these, bringing it into a single pipeline structure, a single option to choose from, and that will allow us to do more automated governance, get people through the journey with push button experience. And it's a big drive. We've got a lot of legacy. As a bank, we've done a lot of mergers through multiple different banks that have come under the banner of Lloyds Banking Group, and we are now really seriously looking at these and trying to give a coordinated single view of how engineering should be done at the bank with the hope that it will help with engineering portability as well. So when an engineer wants to move from Team A to Team B, it won't feel like joining a brand new organization. There'll be familiarity which will increase their productivity there as well.
BP I think it's really interesting to say that some of what is going to be important in 2025 is seeing who can take advantage of some of these new tools and capabilities that are available and use that to broaden their skillset to be able to create things faster, even if they're not themselves a developer or a quant or an artist, but they now have the ability to make a visualization to show ROI through charts and through data analysis and to use tools that kind of empower them with these new skills. Of course, I think it's really important to note all of that is fun, but within an organization like yours, you have to be extremely careful. What am I feeding into this tool to try to get this data analysis, and is this a tool I trust?
TK I think there's a real opportunity that we can see the base level of skills go up significantly, because the access to the greater knowledge of the world all over is available at your fingertips. And I think if you use it in the right way, the speed of learning, the breadth of learning can really accelerate. So there's absolutely opportunity that we would want to capture. Making sure that's targeted in the right way and making sure that we still value real quality skills is absolutely paramount too, and there's going to be an interesting point that we reach where you have to decipher between what is real skill, what's real knowledge, what's learned experience, and what is AI. I don't think the industry or anyone is anywhere near working out how to decipher between those. I see recruiters complaining about this a lot, about how AI is now generating CVs and they're using AI models to analyze those CVs, so it's just a back and forward between two models. And actually what's interesting is you lose that kind of human purity, the human creativity in that process. So there are organizations like ours but like Stack Overflow with Prashanth are going to face into these challenges of deciding what skills you really value, what human skills you really value. I think Prashanth plays into this a lot when he talks about the community value of Stack Overflow and that being supplemented and enhanced by AI, but not the other way around.
BP Yeah, I think that's a great point and a good one to end on. If you're at uni and you're a professor, you're dealing with plagiarism, and if you're a manager in a business now you might not know if that work output was a hundred percent your employee or maybe AI, and it's probably worth asking them what kind of tooling are you using and how much of this did you come up with yourself? And if they're able to do great things with AI tools, that can also be useful too.
[music plays]
BP All right, everybody. As always when we get to this part of the show, we want to shout out someone who came on Stack Overflow and helped us out by sharing a little bit of their curiosity or their knowledge. A Populist Badge was awarded yesterday to Andreas Bergström: “How to pass a message from HttpResponseRedirect in Django?” Andreas gave an answer that was so good it got twice the score of the accepted answer, and so they earned themselves a Populist Badge. Congrats, Andreas. As always, I am Ben Popper. I'm one of the hosts of the Stack Overflow Podcast. If you have questions or suggestions for the show, email us, podcast@stackoverflow.com. And if you like what you heard today, do me a favor, leave us a rating and a review.
PC Thanks, Ben, and thank you both Hills and Tom for joining us. Amazing insights. My name's Prashanth Chandrasekar. I'm the CEO of Stack Overflow. You can learn more about all our knowledge solution products and enterprise products at stackoverflow.co.
HL I'm Hilary Lanham. I'm Head of Engineering Experience at Lloyds Banking Group. And if you'd like to know about where to go to find role opportunities for Lloyds Banking Group, please go to www.lloydsbankinggroup.com/careers.
TK And thanks, I'm Tom Kelk. Please feel free to reach out to me on LinkedIn. Otherwise, please go and have a conversation with someone at Lloyds Banking Group about careers.
BP All right, everybody. Thanks so much for listening, and we will talk to you soon.
[outro music plays]