What happens when access to advanced AI models is no longer the real differentiator, and the true advantage lies in how businesses leverage their own data?
At the AWS Summit in London, I sat down with Rahul Pathak, Vice President of Data and AI Go-to-Market at AWS, to unpack this question and explore how organisations are moving beyond experimentation and into large-scale generative AI adoption.
Recorded live on the show floor, this conversation explores how AWS is supporting customers at every layer of their AI journey. From custom silicon innovations like Trainium and Inferentia to scalable services like Bedrock, Q Developer, and SageMaker, AWS is giving businesses the infrastructure, tools, and flexibility to innovate with confidence.
Rahul shared how leading organisations such as BT Group, SAP, and Lonely Planet are already applying these tools to reduce costs, speed up development cycles, and deliver tailored experiences that would have been unthinkable just a few years ago.
A key theme that emerged in our discussion is that data, not just models, is the true foundation of effective AI. Rahul explained why unifying data across silos is critical and how AWS is helping companies create more intelligent applications by connecting what they uniquely know about their business to powerful AI capabilities.
We also addressed the operational realities of AI deployment. From moving proof-of-concept projects into production to meeting the growing demand for responsible AI, the challenges are shifting. Organisations are now focused on trust, security, transparency, and measurable value.
If you're leading digital transformation and wondering how to scale AI solutions that deliver on business outcomes, this episode provides practical insight from someone at the center of the industry. How will your business stand out in a world where every company has access to AI models, but only a few know how to apply them with purpose?
[00:00:04] What happens when generative AI stops being a proof of concept and starts driving real business outcomes? Outcomes with a measurable impact. That's a question I brought to today's guest as we sat down together at the buzzing show floor at the AWS Summit here in London.
[00:00:23] His name is Rahul Pathak. He's the VP of Data and AI Go-to-Market Strategy at AWS, a role that places him right at the intersection of enterprise ambition and AI execution. And at a time when organisations all around the world are moving beyond experimentation and beginning to embed AI into the core of their operations,
[00:00:47] I wanted to get his take on what it really takes to move from hype to meaningful, scalable adoption. So in our conversation today, we're going to cut through the noise and explore how AWS is helping customers from BT Group, SAP or Lonely Planet and help them all leverage AI as a practical toolkit rather than one single silver bullet.
[00:01:11] And my guest will explain why access to powerful models is no longer enough and how the real competitive edge comes from combining that power with your own unique data. And he'll also share insights into agent-based AI, sustainable innovation through silicon like Tranium and how AWS is approaching responsible AI design and deployment.
[00:01:37] But this isn't just a technical conversation. We're going to talk about the human side too. The developers are building these systems, the leaders making the tough decisions at the helm and the end users expecting personalised frictionless digital experiences as standard. So what does it take to go from pilot to production with AI? And are you already behind if you haven't started?
[00:02:02] With that scene perfectly set, I'm going to beam your ears directly onto the show floor here at AWS, here at the AWS Summit in London. So thank you for joining me here on the podcast here at AWS Summit live on the show floor in a nice little booth. But for everyone listening, can you just tell everyone a little about who you are and what you do? Absolutely. It's great to be here. Thanks for having me. And I'm Rahul Pathak. I'm the VP of Data and AI Go-to-Market Strategy for AWS.
[00:02:31] And that's basically working with our customers and partners and helping them make sure that they're successful in achieving their business objectives by leveraging generative AI. And there is so much focus on technology and predictably AI, but I think the real beauty of the summit is just bouncing ideas around people that you don't normally get to meet. And it's that human element. And there is so much happening with AI that can be difficult to keep up.
[00:02:57] So can you start by just giving us a high level overview of AWS's approach to generative AI and ultimately how you're helping customers move from that experimental phase to full scale adoption? Because that's where we are now. We're three years in, right? Absolutely. And, you know, AWS is investing comprehensively in AI. And our belief is that every single aspect of the customer experience that we're used to engaging with will be impacted by AI.
[00:03:23] And so we're really trying to invest at every layer in terms of infrastructure, networking, silicon services needed for training like SageMaker HyperPod, for inference like Bedrock. And then we're also working to provide customers that want a higher level of abstraction access to intelligent assistance that we provide through Q and then also providing infrastructure for building agents and agentic applications. And so the goal is to really help customers get whatever they need to be successful with AI.
[00:03:53] The other piece that we're really big believers in is that in order to differentiate what you're doing with AI, you really need to think about how you connect to data. Yeah. And ultimately, that unique data coupled with state of the art AI is what's going to help customers succeed in production. And we're really committed to helping customers do that. And I'm so glad you said that because I was reading before we you joined me on the podcast today that you've said that access to powerful models on their own is no longer the big differentiator. It's all about the data.
[00:04:22] So what is it that makes data the defining factor in that successful AI implementation, that thing that everybody's chasing right now? Well, if you think about it, anybody with a credit card can access state of the art AI today. So because anybody can do that, there's nothing unique about that capability. And that's why while models are important and you want access to the best models, which we provide through bedrock, it's not enough.
[00:04:45] What you need to do is to be able to combine what you uniquely know about your business and your customers with state of the art models. And then you can develop AI experiences that are uniquely tailored to deliver value for your customers in a way that only you can accomplish. And that's what we mean by it's actually data plus AI is how you differentiate. And there was a great moment in the keynote today talking about data. And obviously, for most organizations, they don't have that ready-made data. It's silo data. There's things all over the place.
[00:05:15] Could you expand on that for anyone that didn't see the keynote and how you can unify that data? Absolutely. So, you know, we've thought a lot and have really invested over the years in making it easy for customers to work with data. So one of the capabilities that we offer is SageMaker Unified Studio. And the goal there is to allow customers to really catalog and access and query and combine all of their data assets and then make that data available to then serve analytic and ML and AI applications.
[00:05:43] And so it's that ability to connect to data wherever it happens to live, bring it into one place and then work with it and use that to provide the underlying foundation for the intelligent applications that customers are trying to bring that I think is really powerful. And as we said at the beginning of our conversation, we're now in the third year since generative AI and the world went crazy. And a lot of people just rushed into things and went tech first and not problem first.
[00:06:08] So this year, there seems to be an increasing focus on ROI value, measurable difference for every single tech project out there. So is there anything you can share around some of your high profile AWS customers? And there's some big names there from Dentsu, SAP to BT Group. How are these companies already applying your AI tools like Amazon Nova, Amazon Q developer to achieve these meaningful results that we're talking about and every business is chasing?
[00:06:35] One of the things that all of our successful companies share is sort of a mindset when it comes to operating with AI. And really, this starts with working backwards from the customer outcome they're trying to achieve. So what is the business objective? Are you trying to make more money? Are you trying to save money? Are you trying to increase your time to market? Are you trying to speed up your time to market? And once you've got those business objectives, then thinking about how AI can be one of a suite of tools that you use to achieve those goals is really what all our successful customers share.
[00:07:05] And so folks like BT, they're having, you know, a large percentage of their code being auto generated by Q developer, and that's speeding up and improving developer productivity. Dentsu is working with Amazon Nova to generate advertising creative, and that's using our image generation, video generation models. And so, you know, really what we're trying to do is provide a broad suite of capabilities and then use those to help customers deliver on the objectives that matter to their business.
[00:07:31] And to your point, it's really all about thinking about what is the business outcome we're striving for, and then how do we most effectively apply AI as one of the tools available to us to kind of achieve that objective. And there's a word we've used a few times here, tools. And rather than a solution or a service, it really seems to be this selection of tools for almost any scenario, right? So, AI is incredibly versatile. And, you know, I think the places you can apply it are really limited only by our imagination.
[00:08:02] Yeah. And ultimately, you know, the way to think about the technology is that it's incredibly powerful. It allows us to do a lot of things, but you don't want to start off with the technology looking for a problem. You want to start off with an objective and use this as a force multiplier as a way to get you into that problem. And what AI allows us to do is to, you know, frankly, do things now that seem like magic six months ago or a year ago.
[00:08:24] There's so much innovation that's happening and putting that in service of what we're trying to accomplish to improve an end customer outcome, to save time, to reduce the amount of time we spend on overhead. Those are all things that are really transformational. And it's a very exciting time to be in tech at the moment. It really is. And as you said, six months, what you thought six months ago would have been magic is now reality. And although it's moving at such a fast, fast pace, there's also an argument that it will never move this slow again.
[00:08:53] And where are we going to be in another six months? What do you find most exciting about being right in the middle of this space and being in the eye of the storm with so much change? Well, what's amazing is, you know, I think this technology is going to be everywhere, but it's only being born now. Yeah. The fact that we get to participate with our customers in inventing and in shaping this and in seeing it take hold is incredibly exciting. And we work with some amazing customers and partners to try and make this happen.
[00:09:22] I'd say the other piece that's super exciting about this is that, you know, I think we're really just starting to scratch the surface of all the possible applications that we can see. And I think as customers are using AI today and as they're thinking about how to use agents and agentic workflows tomorrow, we're going to just see a massive explosion in applications and productivity. And it's going to be, I think, an incredibly exciting time.
[00:09:46] And then on a personal level, I'm sure you're seeing this too, but, you know, almost all of human knowledge is summarized in these AI models and in these AI capabilities. And so being able to plug into these with the ease that we have today is also remarkable. It's just from a personal learning perspective. It's like plugging into the matrix without all the messy cabling. So it's incredibly exciting. I'm still waiting to wake up and know how to do Kung Fu. That would be great.
[00:10:13] But, of course, anyone that's been involved in any tech project will know there's a few problems and challenges and obstacles. So I'm curious, from all the conversations you've had, what are some of the biggest challenges that you've seen enterprises coming across, especially when scaling generative AI? And how are you helping them with AWS? How are you helping them overcome some of these challenges? One of the challenges that we see is customers often struggle with how to get from great POCs into production at scale.
[00:10:47] Yeah. Operational excellence and then the price performance at scale that you need.
[00:11:12] And that's an area where I think AWS has been really at the forefront in helping customers think through what's your choice of models? Are you going to operate at scale? What is the right price performance profile? Which is the best model for which use case? How do you securely connect to data? And we've really worked hard to make all of that easily accessible to customers through the services that we provide. And something else I wanted to bring up is that AWS has made some pretty big moves in custom silicon with Tranium.
[00:11:41] How are innovations like Tranium 2, Tranium 3, how are they changing the equation in terms of AI performance and cost efficiency? Something right at the heart of every IT strategy, I suspect. Absolutely. So we've had a long history of working with custom silicon and with Tranium and Inferentia. We're trying to design silicon that delivers great price performance for customers' most demanding training and inference workloads. And what we're seeing is customers can save up to 40% from a price performance basis.
[00:12:10] And what that in turn translates to is a better cost, better return. It also importantly means better energy consumption, so it's more sustainable, which is really important as we scale AI. And so we're working closely with our customers and our partners and even folks such as Anthropic who are training their latest models on Tranium 2 to give this as an option for customers so that they can get the most out of their investments in this space.
[00:12:36] And you mentioned sustainability there, and obviously there's some very highlighted problems around sustainability in AI. And there's so many big responsibilities that come with AI, and responsible AI as a result is becoming a huge focus. So what steps are you at AWS taking to ensure AI deployments are safe, transparent, trustworthy, and sustainable? We're investing a ton in features around responsible AI, and so we have a ton of guardrails built into the things that we offer through Bedrock.
[00:13:05] We're trying to make sure that we give customers the features that they need in order to be able to ground the prompts and ground the responses that they get so that they're focused and on topic. We also provide transparency around the models and the model cards and what their strengths are and what use cases they're best suited for.
[00:13:23] And then we provide some advanced capabilities like automated reasoning, which is something that no other cloud provider offers, where you can actually define a policy, and then the system will prove that your responses comply with that policy. So it gives customers a lot of confidence that the responses they're getting are aligned with the problems they're trying to solve and not going off topic. Here we are right in the heart of the show floor, and there's so many different conversations.
[00:13:50] One conversation I keep hearing again and again is, of course, agentic AI and bots and agents, etc. What are you hearing or what trends are you seeing around agentic AI and indeed any other trends that excite you right now? So agents are a big part of the conversation we're having with customers and a big part of where we're investing as AWS. And our vision is to really provide the most secure and most scalable infrastructure for customers to build agentic apps.
[00:14:17] And so we have agents that are prepackaged, for example, in Q Business or Q Developer that allow customers to interact with enterprise data or to build applications or to cover the full software development lifecycle. So that's out-of-the-box capability. And then we're investing in infrastructure so customers can build on top of that using managed agents from things like Bedrock with Bedrock agents or working with the latest in open source capabilities.
[00:14:42] Because, again, the space is moving so fast that we want to make sure that we're providing a very open environment for customers to build. And it's an incredibly exciting time. I think we're going to see a ton of automation, ton of innovation. And I think it's going to be a real force multiplier for our customers and enable them to get a lot more done with the teams they have in place than they do today. And there's an old saying that the last best experience any of us have anywhere becomes the standard expectation for what you expect everywhere.
[00:15:10] This podcast booth is a great one for me. I'm going to expect this at every event from now on. And you've mentioned a growing trend from internal to external AI applications. So anything you can share on how AWS customers are starting to build customer-facing AI experiences? Because they're going to set a new expectation right there. Anything you can share around there? We're seeing rapid adoption of customer-facing AI. And this is in a number of places. So I think a lot of customers started with internal use cases because they're easy to get up and running.
[00:15:40] They're in an environment where you can debug and deal with issues as they come up without impacting customer experiences. But as our customers have gotten more confident, they've started to scale these out externally. So you start to see it in customer support. Customers like DoorDash, they have their support agents actually work with AI and directly interact with customers and automate a number of those support engagements that they have before they ever get to people. We've got customers like Lonely Planet.
[00:16:08] I don't know if you ever looked at their travel guides in the past. Yeah, got me out of a few scrapes in my time. But now with AI, they're actually able to generate custom personalized travel guides for each individual based on their preferences. Wow. On the fly for pennies, something that was completely out of reach before. And now based on your profile, you can get a guide that's custom tailored for you and the things you like to do based on all of the knowledge that Lonely Planet has about places to visit.
[00:16:34] And so these unique custom experiences that can be generated on the fly are all over the place. And so it's, again, I think really unique in terms of driving engagement within customers. It's such an interesting space at the moment. My son recently did a digital nomad thing. He found himself in the middle of Turkey. He didn't know which bus, which train or how to get out of there. He said he went to Google and the first 10 results were sponsored results that helped him in no way. So he went to AI, got all the information he'd instantly.
[00:17:03] So even things like search is really up for debate and changing right now, isn't it? How we search for any information. I do think all of our modes of interaction with information are we're going to see innovation that's AI driven. And I think the ability to ask a focused question that's targeted at your particular need and get back a response that's tailored to your specific question, that's incredibly powerful. Yeah. And I do think we're going to see this change continue.
[00:17:32] And, of course, there are many business leaders listening around the world. They're still sitting on the fence. Maybe it's fear of regulation further on down the line or just unsure how it's going to play out. But I think one thing we can all agree on, it's not going anywhere. So for any business leader listening who is still cautious about diving into generative AI, maybe they have still banned it in the workplace. What's your message to them? And what might they risk by waiting too long? My message to them is very simple. It's if you're not leaning in, you're going to be left behind.
[00:18:01] And the rate of change in this space is so fast. And customers are in production today. Customers are actively putting this technology to work. And humans are really bad at understanding exponential change. And I think if you're not getting started today, your peers are. And you're going to look back in six months and really not like your position. And so our advice to customers is let's get started. Let's cut our teeth if you haven't already.
[00:18:29] And the best time to start might have been a year ago. Second best time is today. So let's get going. And you've had an incredibly long flight to attend the AWS Summit. And you've spoke to so many different people on the show floor watching keynotes and interactive sessions. When you get that long flight home, what are you going to be thinking about? What are you going to be reflecting on from all those conversations that you've had? Anything going to be taking away from this event? It's been super energizing to be here.
[00:18:57] There's so much innovation happening across the world. But specifically in Europe, there's a ton of innovation happening in the AI space. And I've met just a ton of great customers. And I'm going to be looking forward to following up with them and figuring out how we can grow together and innovate together in the months to come. Fantastic. Well, thank you so much for sitting down with me today. Before I let you go, obviously, we've been talking about the pace of change.
[00:19:20] Where's the best place for people listening and watching to stay up to speed with the kind of work that you're doing and find out more information? I would just stay connected to our blogs and aws.amazon.com and reach out to a local AWS solution architect or contact your account manager. But we're very excited to help all of our customers stay up to date and innovate together with us. Fantastic. Well, I know how incredibly busy you are. But we've covered so much in a short amount of time.
[00:19:49] I'd love to hear from people listening and watching and thought what they take away from this, especially anyone attending the AWS Summit. But more than anything, thanks for sitting down with me today and starting this conversation. Thanks so much for having me. It's been a pleasure. So a big thank you to my guests and the event organizers here at the AWS Summit for creating this incredible podcast booth. It did feel like I was in a zoo during some parts of the interview because as we're talking, obviously there's see-through windows and people are walking around taking photos.
[00:20:17] So I had to try and keep my game face on. But it felt like a real game changer being able to record an interview like that right in the heart of the show floor. So kudos to everyone involved for making that happen. And as our time here on the show floor wraps up, one thing that feels increasingly clear is that AI isn't just another layer of enterprise tech. It's becoming part of the fabric of how businesses operate, compete and serve their customers.
[00:20:47] And if you're still thinking about whether you should get started, the moment to act might already be passing you by. Because I think Rahul's perspective today really served as a timely reminder that real competitive advantage lies not in AI itself. Let's dispel that myth straight away. It's in what you do with it, how you apply it, the data that you pair with it and the outcomes that you choose to aim for.
[00:21:13] So whether it is using generative AI tools to speed up software development, redefine customer support or offer personalized experiences at scale, there are a variety of tools there to help you do that. AI isn't that silver bullet that does everything. It's in those variety of tools at your disposal. And the only question that remains is how fast and how will you use those tools? So if you are a business leader still weighing up your options, still sitting on the fence, what is it that's holding you back?
[00:21:43] What problem could you be starting to solve right here, right now? Even if it's with a simple pilot, one that sets your team on the right course, it's better to do something than nothing. And I'd love to hear your thoughts here, especially if you were also at the AWS Summit. Maybe you were one of the people walking around the podcast booth taking photos and you've stumbled across the podcast. Whatever it is, I want to hear from you. This is a dialogue, not a monologue. So please drop me a message or join the discussion.
[00:22:11] Just keep this conversation going. Techblogwriteroutlook.com, LinkedIn, X, just at Neil C. Hughes. And remember, if you are still hungry for more content, check out techtalksnetwork.com. I now have eight individual podcasts focusing on niche areas that will allow you or your enterprise to keep up to speed with developments out there. So I will speak with you all very soon. Thank you for listening. Bye for now.