Cloud Strategy in the Age of AI: Navigating Nuances and Opportunities

Speaker: Cloud strategy in the age of artificial intelligence. Let's talk about navigating nuances and opportunities. Please welcome to the stage, Chris Aniszczyk, who is CTO at Cloud Native Computing Foundation. Allison Ortiz, who is WW industry head at GSI, Infosys, and AWS, and Haja Deen, who is global data and analytics director at Pladis Global deliberated on the topic and moderated Anand from Infosys.

Anand Ramakrishnan: Good morning, everyone. My name is Anand Ramakrishnan and I'm from the consumer goods retail and logistics vertical of Infosys. Today we've got a topic that is very, very relevant for all of us. Cloud has been making a lot of strides in the last few years and it's only accelerated during the pandemic as well. AI is actually gaining more and more prominence and pretty much we spoke about some of those topics in the previous panel as well. So what we're here to discuss today is to share some of the experiences of the distinguished panel that I have here with you. So Haja is an industry leader from Pladis. He leads the data and analytics practice at Pladis and Pladis is a company I think all of us are familiar with their brands, whether it's Godiva, McVities, or Jacobs, just to name a few.

Chris, welcome Chris, is part of the Cloud Native Computing Foundation. He's the CTO and he's leading the cloud-native development, especially towards all the technologies that we see today. They also have projects such as Kubernetes and Prometheus, which I think some of you would be familiar with.

And Allison is the global partner leader from AWS. She leads the strategy and the solutions for this practice and is working closely with Infosys to see how we can add value. So welcome everyone.

Chris, let me start with you. So as someone at the forefront of cloud-native development, you've seen a lot of challenges in the industry. You've seen a lot of developments. So how do you see this entire evolution of AI and its relevance to the cloud today?

Chris Aniszczyk: Yeah, it's a little bit crazy where, you know, as our organization, I don't know how many are familiar with it, but we're all about kind of democratizing and open sourcing technology that essentially, you know, the Internet pioneers, the Googles, Amazons of the world have used and kind of making that available for everyone to kind of build upon. And in the last few years have been a little bit crazy for us, where obviously the pandemic caused an acceleration towards cloud usage. We got crazy busy during that time. This kind of resurgence of, you know, modern, you know, AI that's kind of coming up all over the place is causing us to be incredibly, incredibly, incredibly busy. I see, you know, companies within our ecosystem all looking to kind of improve their digital transformation efforts of seeing, you know, what can we learn from companies in the cloud native ecosystem and, you know, kind of, you know, improve things. But what's, you know, in terms of like trends, which I think your question is really about is, you know, I don't know how many people are here familiar with like the concept of, you know, Shadow IT, right, you know, developers just or, you know, people in your company just installing stuff. So, you know, a lot of times we're talking to people like, hey, I don't know if our like company is really like using any AI stuff now or who knows, like completely false. Like I would go talk to your rank and file employees or developers, you know, people are using stuff, right, whether it's, you know, developers that have installed copilot or some kind of AI-assisted tools to make their lives easier because developers are inherently a little bit lazy, you know, sometimes or your maybe your marketing or PR teams, you know, hey, I need something that will help with like copy or, you know, increasing on the content that I'm created. So it is being used all over. I think you kind of, you know, I don't know if you want to call it shadow AI, but talk to the folks within your actual, you know, enterprise. It's being used. You kind of have to kind of figure out maybe how to manage it like we had to do with the previous generation of IT tools. But it's truly like increasing cloud adoption, you know, for sure. And a lot of the changes are happening extremely quickly where even myself, who kind of works at the forefront of technology, find it a little bit difficult sometimes to keep up.

Anand Ramakrishnan: When we were talking earlier, I was telling you how I was able to write a rap song using AI and then and then build it through the voice of Eminem through, you know, text-to-speech. So a lot of fun out there.

Chris Aniszczyk: It's incredible. Yeah.

Anand Ramakrishnan: Allison, AWS has been making huge investments in AI as well. So how do you see the convergence of AI and of course cloud? And what are some of these stories that you can talk about?

Allison Ortiz: Well, we've had quite a lot of announcements recently, but you know, it starts with cloud and Uma even talked about it at the very beginning, which is really about allowing for the growth that we're seeing. So, when we think about AI and ML, it's not new to us, right? We've all shopped, we've all experienced the customer recommendations, and the product recommendations through Amazon.com in our retail site, as well as our fulfillment. So now that we've ordered all of these fabulous things, how do we get them to you faster, better, and more accurately, leveraging the AI and ML technologies within our fulfillment centers for our robots? We think about Alexa devices that are in every room in my home and how she's continually learning to create a better experience for you and then completing that complete customer journey. So with all of that comes a lot of data and a lot of throughputs and what we can do with that. So AWS has recently announced a couple of new chips within our EC2 instances to expedite that and allow for that data volume. You know, we think about all of the technologies around that, whether it's 5G to accelerate and reduce latency with all of that data so that it's more real-term and real-time and faster for you. So most recently we announced a hundred million dollar investment into a Gen AI Innovation Center and that's going to bring all of our data scientists and experts to our customers to bring that to them and to your point, right, to democratize AI and make it accessible to our customers, both current and future. So there's a lot we can do. You know, you talk about lazy developers and I'm sure some of us, most of us have probably been that at some point in our careers and we've developed Code Whisperer, which instead of developers having to go through a search online to find a snippet of code that's going to get them 70% there or 80% there and then they can layer in the rest of the personalization that they need to make it their own and fit the needs of their business, Code Whisperer does that for you. Bedrock is introducing the foundational models to go through and allow for the use of APIs to get to those learning models for you faster. So there's a lot we're doing. I'm really excited about the Innovation Center, certainly, that's going to help accelerate the growth.

Anand Ramakrishnan: That's wonderful, Allison. Haja, you've been busy setting the foundation of your landscape, especially around data and analytics for the last few years. In today's inflationary environment, costs going down, spending going down, it is extremely important for you to look at what you need from an insights perspective. I know you call BI to AI as part of your own transformation, so could you just talk about it?

Haja Deen: Yeah. Thank you, Anand. So many of you may not have heard of Pladis, but you would have heard of McVities or Jacobs or whatever one of the brands we are fortunate to be custodians of. We started our analytics journey around two years ago and it was by default we were going to do it on the cloud. We partnered with Infosys to get us started on this transformation. In that situation, there's been a lot of macro factors as the previous speaker alluded to, so high inflation and we couldn't really predict what the consumer was going to do next. All of our historical data, what it was telling us, and what we were seeing happen in terms of our pricing or promotion effectiveness. To put it mildly, it was a bit in disarray, I would say. We had to react quickly and sort of rebuild models. Some of it was just on Excel, and some of it was machine learning. We had to rebuild these models to get a better grip on what actually was going to happen when we changed prices or when we changed the size of our products or even ingredients. That mattered a lot to us because our business motto is actually happiness with every bite. As part of that is being connected to the consumer, we sell our products almost in every country in the world to make sure that we can still serve our consumers in a meaningful way while taking care of all these inflationary pressures which was happening. With the tools we had from both AWS and our Infosys partnership, we were able to build some models quickly tested in the business and last month we delivered a complete set of revenue growth management tools into our North American business which completely increased the accuracy of our predictions. We could have only done that because we were already based on the cloud and we were thinking in an agile way. This is again playing a big role when it comes to now the generative AI space and how we are planning to use that in almost every aspect of our operations and planning.

Anand Ramakrishnan: That's amazing. Maybe as a follow-up to that, data is critical to all of this decision-making. You can only take decisions on insights that you get. You must have had lots of pushback and challenges from the business as well. So how do you deal with some of those unknowns and especially challenges that you will encounter? What are the learnings?

Haja Deen: Great question Anand, the one that we ponder over every day. Because in a business-like Pladis, we have around 16,000 colleagues around the world. Making great snacks is a bit of a science and an art. There's a lot of data involved. We produce data around the world but it's also an art and we call ourselves at the heart of it, we are bakers. So there's that intersection between science, data, and art, we need to sort of tread very carefully and with great respect to colleagues involved. We don't really think of it as change management. However, the mantra for the analytics team or the digital team which I lead is the solution we put in must first be painkillers, must solve some issues or challenges for the colleagues, and then it must be vitamins. So it must make their lives better. The technology is secondary. Technology is a tool that has become significantly better over the last 20 years that helps us to do this both the painkiller part and the vitamin part. When we do that, what we realize is that the change management sort of takes care of itself. Obviously, not everybody will come on that journey, you have to deal with that. But change management then becomes easier. The pushback becomes a lot easier to manage. As an example, I was speaking to one of our senior North American colleagues two weeks back and he is not technical. He was able to articulate how this AI-based revenue growth management tool we have put into his business was significantly better than everything else he has seen in 25 years. To us, that's change management taking care of itself.

Anand Ramakrishnan: Amazing. Allison, I know there are plenty of use cases you can talk about. Why don't you give us a few examples in this space as well?

Allison Ortiz: Sure. Well, I touched briefly on the development and Code Whisperer to accelerate that because when we think about the way customers interact with us and what Amazon is doing on behalf of our customers, it's really to remove friction and to improve the overall customer experience and allow for value-added activities. Remove the work and optimize the effort that's needed for non-differentiated actions and create that personalized experience. Some of the work that we're doing with Infosys, in particular, is around the contact center intelligence solution, where we're taking advantage of a couple of our solutions within Transcribe and Extract Edge and Textract to take the non-digital work and make it digital, to take the data from documents and move it into our digital assets, and then to create insights on all of that, so to optimize that. In some of the areas that we're working on, Infosys is a launch partner for Bedrock and for Code Whisperer. Publicly, we've been recognized together with our different speakers and solutions through our press releases and blogs, and so on to highlight that partnership because we're always going to work backwards from our customers and innovate on their behalf. Our team meets regularly to think about what's going to be innovative and what's going to be differentiating for the industries. We look at cross-sections, and cross-verticals, and see what we can do together. I think with that, where we can do and continue to innovate, that's where we need everybody's help.

Anand Ramakrishnan: Thank you. Chris, open source is at the center of cloud-native architecture and adoption, and the general perception is, oh my God, it's so complicated. It's going to cause me delays in development and operations and all that. How should companies deal with this, is it a myth, and how do you deal with some of these factors going forward?

Chris Aniszczyk: Cloud-native in general, anytime you have distributed systems, you inherently have complexity. Things fail. How do you deal with failure? How do you remediate things? That problem's never going to go away. I think if you're new to this, finding the right partners, whether it's large cloud providers, or Infosys of the world, working with them to get you to the next level will definitely help, but this stuff isn't easy. You're not just running on one little server now. You have complex systems, lots of services, all intertwined, talking to each other with crazy failure modes. I think every kind of company does it a little bit differently. I think also about the willingness to learn from peers in the industry and potential competitors. Our open source foundation, for example, hosts 850 or so companies that all work together, sometimes compete in the same industry, sometimes collaborate across industries, but it's all about learning and sharing knowledge of how maybe they build out a platform engineering team, or how do they use Kubernetes in an air-gapped environment or something. The willingness to work and learn from folks in history I think will help you a lot, especially if you have the right partner that got you. I've seen examples all across the board. I'm trying to think of a funny kind of story to share that's kind of relevant to the theme of cloud and AI today. We hosted a conference in Amsterdam recently, which is one of the largest open-source conferences in the world for the Kubernetes and cloud-native community, and I caught up with a company in our ecosystem. I don't know if people here are familiar with the organization called Chick-fil-A. They're kind of like a fast food joint in the US, but you would not think of them as a bleeding-edge technology company. They're in the business of making fried chicken sandwiches and delivering them to people, but they've done a great job of kind of embracing modern cloud-native practices and open-source technologies. They have something they kind of dubbed enterprise restaurant compute, which is kind of a funny name for me, but they essentially treat each fast food retail store as an edge piece of there. They've deployed modern cloud-native techniques, which basically allow them to have a consistent experience across all their stores, but also kind of do what we cloud-native developers like to do. We like to do canary rollouts. Let's go test something in a particular store and see how it does and quickly be able to deal with any failure, and roll back if we need to do that. They've done such a great job that they're starting to play with all this kind of new AI technology out there and are able to test it in some stores, see how it goes, and quickly if something goes bad, they're able to kind of roll back. So I think just learning from folks and folks like that, it just will help you kind of along your kind of cloud-native journey and digital transformation. So there are countless other kinds of examples, but that one just kind of stuck in my head this morning.

Anand Ramakrishnan: Wonderful. Haja, so you've set the foundation, you've had small wins, and you want to scale it out. So what are some of the things that people here can learn from in terms of making it successful? Are there organizational changes? You briefly touched on the change management part of it. So, what should companies do to be successful in the journey?

Haja Deen: Great question, Anand. The way we think about it is, we create a strong vision and we establish that as the sort of the guiding light. In fact, Amazon's e-commerce is one of our strong visions and actually, is what guides us. We sell this idea into a business that looks like you order a product on Amazon, and nobody touches it. It's all completely digital from the ordering part to the fulfillment, everything in between pricing, and promotion, everything is taken care of by algorithms. So there will come a day, and internally we have a small sort of a prediction around this, there will come a day more CPGs or manufacturers like us will get to that state, where the office-based functions will be run by algorithms. So that's our kind of North Star. Now, clearly, we are not there yet, nowhere near it. So all of our effort and activities, POCs, and scale-ups are all directionally right towards that. So, when we run POCs or when we get funding for POCs, we sort of sense check, is it heading us towards this vision where we can pretty much run the office without people? That's kind of a utopian view, there will still be people involved, but directionally we are moving towards that. Small wins are necessary to create a belief within the organization that technology can help us to do this. And interestingly enough, with this new sort of buzz around generative AI and AI in general, it's actually helping our cause that it's now taken for granted. Nobody would question this within the organization that there will come this day when we will be able to run most of our current activities using AI. So that's sort of giving us a bit of a tailwind in creating, converting some of our small wins into bigger wins, and we sort of label it from BI to AI is a kind of headline we give it to move in that direction.

Anand Ramakrishnan: That's a very ambitious plan indeed, thanks Haja. Allison, you spoke about the investments that AWS is making and all the great platforms that you've built. So what should people here do to be successful and to engage in some of the wonderful things that you're investing in?

Allison Ortiz: Yeah, so take advantage of our innovation center, first of all. I think that that's something that learn from the experts, but if you don't have a data strategy today, most of us do, but if you don't create one, if you do revisit it and understand and identify what is the framework for the use cases that are going to be important to you, and what are your non-negotiables? What can you not live without? And then create a lot of those use cases, leveraging the different areas of the business and bringing those together, because that's where you're going to see that overlap in the efficiencies of what you can do together. Get started and create those POCs and fund them start with three and find what works and pivot where you need to. Leverage our experts, leverage the expertise of Infosys, and rinse and repeat. So, I think the biggest thing is to get started.

Anand Ramakrishnan: Thanks, Allison. I've got the one-minute board up. So the last question for Chris. So, you spoke about having the right partner, the right solutions, the right technology. Do some future-gazing? What do you see in a few years from now in the space?

Chris Aniszczyk: Yeah, so quick one, I thought, I think the pace of technology is changing so fast. I heard recently, that from the day the web was created to the iPhone release was like 15, a little over 15 years from iPhone to essentially where we are today, it's been another 15 years. And if I look at that first 15, it kind of felt a little slow, right? The last 15, it's been crazy. And then if you kind of look at from an AI perspective, open sources and impacting AI so much, like in early March, there was kind of a sharing from Facebook about opening some of its models that they had. And in those three months, traditionally, you have to use a very large cloud or kind of centralized thing to take advantage of AI. People have like, the innovation happening open source community have whittled it down so you could run these models on your phones now, Raspberry Pis, all these kind of crazy use cases are happening where you kind of build very specific models for the use. So I would definitely, change is happening extremely fast. Pay attention to what's out there. Look at the innovation that's happening in open communities and find partners to help kind of guide you through there because the change of pace is happening super quickly and it's going to impact us all and if your company is not paying attention, you kind of may lose out on a lot of business in the future. Thank you.

Anand Ramakrishnan: Thank you, everyone. I hope you all had some good insights and ideas on what to do in your respective journeys. Enjoy the rest of your day and thank you once again, Allison, Chris, and Haja.