Quality at Speed: LPL Financial’s Approach to AI-Driven Software Development

Chad Watt:

Welcome to Ahead in the Cloud, where business leaders share what they've learned on their technology journey. I'm Chad Watt, Infosys Knowledge Institute researcher and writer. Today, I am speaking with Mamatha Pathipati, Senior Vice President of Technology at LPL Financial. LPL Financial is the largest independent broker dealer in the US. It supports 29,000 financial advisors and holds $1.7 trillion in assets. Welcome, Mamatha.

Mamatha Pathipati:

Hey, thank you, Chad. Happy to be here.

Chad Watt:

Great. You have worked in financial technology for 20 years. Which was your first love, technology or financial services?

Mamatha Pathipati:

It's the technology that's my first love and the financial services became a perfect partner for me to bring that love to life.

Chad Watt:

Now, Mamatha, do you like to break things?

Mamatha Pathipati:

Of course, absolutely. I love breaking things, but not the way you might be thinking. In my role as a quality engineering leader, my job is to break systems intentionally. What it means, it's finding those weak spots, uncovering those vulnerabilities, identifying those gaps before they can reach our customers. And this is where every test we perform, every bug we found, every failure we detect will help the systems to be more resilient. And this is where me and my team take pride. And that's why I really like breaking things in a different way.

Mamatha Pathipati (Super):

Senior Vice President of Technology, LPL Financial

Chad Watt:

Why does quality engineering matter in financial service?

Mamatha Pathipati:

Quality engineering is important in every industry, but when it comes to financial industry, it's much more important. Because in financial industry, we're talking about protecting what matters to people, right? Which is their hard-earned money, their dream, and their future. So to put this into a context, if we look into when someone is saving their money for their retirement or building their retirement plans, or planning for their child's education, or building a safety net for their family. They're trusting us, our products, our solutions with their lifetime earnings.

Chad Watt:

Software development is moving faster and in more specialized directions. What does that mean for the discipline of quality engineering?

Mamatha Pathipati:

Oh yeah, that's true. Software development is moving much faster nowadays, especially with the advancement in technology, the frameworks, and even more with the AI today. So what it used to take days to develop is now just getting done in a couple of minutes and couple of hours. And this means for quality engineering, it builds immense pressure because we need to deliver at the same pace or even a higher pace to keep up. So, and it's not just about the pace. Along with that pace, we need to make sure the products that we are building are safer, secure, and reliable. So for QE, the QE role becomes different here, right? So we should be an enabler for that faster development and not a barrier, right? And what it means, right? From a quality engineering standpoint, We need to look at upping our game, changing from all the friends, whether it is people or process or tools. From a people standpoint, bringing in and embracing this new technology and innovative ways to identify the issues early. And from a process and a culture standpoint, shifting left on the testing, embedding testing into starting from the design all the way to the development phase and to the production goal.

Chad Watt:

There's an old coder culture about build it once and throw it over the wall and to QA and have them rush to get it approved. Have you done some things to move past that at LPL?

Mamatha Pathipati:

Build it and throw it over the fence. So that's definitely a topic that's pretty close to my heart because I have seen how damaging this mentality could be, not just from a quality standpoint, but for the entire team morale and collaboration. Where we are at LPO, I'm proud to say we have come a long way from that. But like any other transformation, it's a journey. We're almost there. We have a few more pockets where we need to keep pushing forward. But in LPO, we have done a tremendous improvement on that. So to double-click into how we were able to achieve that, we took a multiple stance at it. One, the very first thing is a culture change. So we have to get everybody aligned to quality as a team sport, not as an individual sport. All players in that have to lean in and play the game well for all of us to win. And it's a shared responsibility. It's not a single person's responsibility. It is everyone's responsibility that includes product, that includes developers, that includes quality engineering. That's a key culture shift we brought within the organization. And because of that, we were able to change quality as being an afterthought to quality from the start. Meaning QE is embedded into the Scrum teams. They are part of their PI plannings, they're part of the story refining, they're part of the sprint planning. They work in collaboration with product and dev from day one, from writing test cases to code reviews to plan the feature to the production, including the ramp up and advisor testing and how they wanted to roll this feature out to the production.

These are all from a culture standpoint, right? And then the processes that we have built in. But then the other one is from a tooling standpoint, because one of the things why people behave the way they behave is because everybody is pushed to the edge when they see the feedback at the very last minute. And that's what creates the chaos and finger pointing.

What we have done at LPL is heavy automation, incorporating continuous integration, continuous deployment and continuous testing pipelines. And because of that, we get the continuous feedback when issues happen instead of finding that in the last minute. So we can avoid the finger pointing and that chaos that goes in. Again, like I said, so this is, we've come a long way, but there's still some pockets we need to work through and we are working through actively, but I'm proud to see where we are now.

Chad Watt:

Terrific, thanks for that. Mamatha, give us your take on artificial intelligence and software development.

Mamatha Pathipati:

AI is changing the world. As I said earlier, what it used to take multiple days for a developer to develop is taking just a couple of minutes, couple of hours for them to finish up. So it's a game changer. It's speeding up the software development. Tools like GitHub Copilot, SonarCube, or Tab9, or Sync Code, Reflective UI, you name it. There are multiple tools out there that helps to not just generate the code or auto-complete the code. They're also helping to catch the bugs, automate the repetitive task in seconds. However, what I would say is the human in the loop is super important. All of this tooling can be an accelerator, but they do not decide this is the right block of code that would be pushed to the production to meet your customer expectations. This is where human in the loop comes into the picture. As the AI tools generates these codes, it is the developer and it is the tester to see, yes, this is the right code that I could deploy or I could embed into my project and it means my security, my stability, my scalability requirements to deliver it to the production. So AI is a helping hand to accelerate most of your repetitive tasks or the simple things that you needed to work through or the monotonous things that you need to work through. But then when it comes to releasing it to production, I'm a strong advocate of human in the loop. And that's what we hear at LPL Financial Follow as well. We use AI to handle all of the heavy lifting, right? Whether it is a Copilot for auto-generating the code or auto-completing the code or the other tooling that we have built in to automate the tests or suggest the code snippets. But we always keep human in the driver's seat. It is our developers and QA are the ones who are reviewing, refining and validating the output to ensure it's not just fast, but it is the safe and secure product that we are delivering to the customers.

Chad Watt:

Now, can you talk a little bit more about AI and testing and quality engineering? It sounds like you're using it throughout the process. How is it different from AI on the front end, automated coding generation Copilot work you're describing?

Mamatha Pathipati:

AI in development and AI in quality engineering, in both fronts, it is trying to help or act as an additional hands of help to accelerate either the developer productivity or the tester productivity. So when you talk about AI in code generation, it's about helping the developers to code faster. When we talk about AI in quality engineering or in testing, it is about helping auto-generate the test cases or bring in the predictive analytics to identify the bugs or anomalies in the code even before the QA comes in and starts hitting the ground and starts testing it. And it is about automating those test cases that are generated. It is about maintaining those test cases' stability, establishing the traceability between the code and the user stories all along.

Chad Watt:

What are the special considerations that come into play when we talk about developing and testing software for financial services.

Mamatha Pathipati:

Financial services industry is a very heavy regulatory compliance driven industry. We have zero tolerance for errors. And because of that, the software development and the testing in the financial services industry is super critical. It's not just about writing a code or testing a scenario. It's about building that trust and building that security and reliability in the products that we deliver and in this highly regulated and high stakes environment. So to double click in detail into it, right? What makes this more unique? The regulations are the ruler. Every line of the code that we write need to be strictly in compliance with those regulation rules. If not, it's going to cause huge fines and even worse could impact the licensing as well. Security, it is non-negotiable for us. Financial systems are the primary targets for the hackers and we have the PII and PCI information in our systems. So it's super critical for us to make sure these systems are secure. No room for error, right? Zero room for error. A tiny bug can lead to my ASO, you know, huge financial losses. It could completely erode the customer trust. We work based on our customer's trust and we have to be reliable and highly performing because our systems are critical systems. We are dealing with moving money, managing money, day in and day out.

Chad Watt:

And that kind of gets to the point of the next question. Financial services companies are more comfortable, it seems, I would say, broadly, with old, tried and true technology, rather than eager to push the envelope or experiment with new technologies. My understanding, that's a little different at LPL. You've been motivated, and you guys have moved to the cutting edge in some areas. What motivated LPL to pursue this?

Mamatha Pathipati:

So in LPL financial, right, our motivation to move to the cutting edge technology was driven by multiple factors. It's a combination of growth that's on our way and we're experiencing already and if you have seen our growth and then the scale and the enhanced service for our clients. So when we talk about the client expectations, right, in the financial services landscape, In the financial services landscape, we're evolving super-fast. Our clients have the same level of expectations on the innovation and convenience like any other tech industry. Whether it is a seamless digital experience or the real-time insights, we knew we had to step up to meet and exceed those expectations. And this is where having in the latest and greatest technologies where we can build in predictive insights and real-time insights for our customers as needed in high-performing systems, that makes it a differentiator for us. So that's on the client expectations. And then, we have competitiveness. So in the financial industry, it's a crowded market. For us to stand out, it means we have to be faster. We have to be smarter, and we have to be more reliable. By adopting these cutting-edge technologies, we are able to streamline our operations, we're able to reduce our time to market, we are able to deliver our solutions that help us set apart and scale. So with growth, scale needed to happen. As we grow, we have to meet the demands on our systems, the number of transactions that happens, the number of advisors that gets onboard and onto the systems. The number of investors that gets onboarded onto the systems and the scale we have to serve with the growth. Legacy technologies will become a bottleneck for us. And hence we realized it long time enough, which led us to move to the latest and greatest technology stack. Last year we migrated to the data foundations and the SORs and that helped us to scale up not just for the growth that we have and for the future growth as well. So again, talking about the futureproofing, right? It's not just about growth but also being future proof. So, embracing this innovation means we are future proofing our business. We're ensuring we're ready for whatever comes next in the fast-changing world, right? And then last but not least, cost savings. With the innovation, with the embracement of the technology, we can deliver faster time to market. And as an example, embracing AI into the entire product development lifecycle. It helps us reduce the operational expense.

Chad Watt:

From a corporate perspective, we're coming out of this sandbox phase with generative AI. How do you and LPL put generative AI to work?

Mamatha Pathipati:

In LPL financial, we have moved beyond experimenting with Gen.AI. So we're now putting it to work in ways that derive real value for our businesses and our clients. So where we are leveraging it, we're leveraging it to enhance our developer productivity. We're leveraging it to improve our customer support. We're leveraging it to, we're incorporating that in the entire product development life cycle so we can be faster time to market and stable and highly, and ensure the high quality with the products that we deliver. And then when we talk about the customer support, we have multiple agents that have put in place which will help automate the back office tasks, which means for a back office, as we grow with the growth that is at us, the back office, we do not need to increase our back office team as these bots and these agents that are put in place would help automate those back office tasks.

Chad Watt:

Just about every cloud platform and SaaS provider will launch an AI agent this year. How do you view the potential of agentic AI?

Mamatha Pathipati:

Agentic AI is going to be very helpful because it's going, it's an outcome driven, right? Instead of being a task driven, so it can help autonomously automate this task by learning, self-learning it. So in LPL financial, we see it as a powerful tool, especially both for our business and for as well as our home office, both advisors and home offices as well. Where we are with the agentic AI, we are exploring right now. We're exploring options to look into how, where we can bring in the agentic AI, whether it is in the resolution center or whether it is helping the advisor be an assistant for the advisor in similar concepts like Jump, right? And then we are also looking into bringing in from the home office customer service representatives, and then the marketing assistance and the compliance team as well. We're looking at all those sectors to see, identify what are the use cases that will have higher benefits, and we're exploring options in 2025 to see what are those use cases that we would be investing in.

Chad Watt:

Thank you, Mamatha, for your time today.

Mamatha Pathipati:

Very welcome, Chad. It's nice talking with you or chatting with you on this podcast. Really enjoyed it. Great questions.

Chad Watt:

Thank you very much. This podcast is presented by Infosys and MIT Tech Review. Visit our content hub at technologyreview.com to learn more. Be sure to follow Ahead in the Cloud wherever you get your podcasts. You can find more details in our show notes and transcripts at Infosys.com/IKI in our podcast section. Thanks to our producers, Christine Calhoun and Yulia De Bari. Dode Bigley is our audio technician. I'm Chad Watt with the Infosys Knowledge Institute, signing off. Until next time, keep learning and keep sharing.