Telefonica Tech Reimagines Itself for the Hybrid Cloud Era
We continue the global tour here at the Microsoft Hybrid Cloud Partner Podcast. I'm Fede Pacheco and this is Adam Burke. How you doing, Adam? Hey, I'm great Fede. How's it going? It's great. Today we're heading to London in the UK with a partner called Telefónica Tech and I'm excited to hear about their transformation story and especially the impact that they're having on their customers. Yeah, we've got Alex and Martyn, both based in London. Looking forward to the conversation here. Thank you, Alex and Martyn, so much for joining us. Thank you for having us. Yeah, no problem. Glad to be here. Would you mind introducing yourselves a little bit? Tell us about your role at Telefonica Tech, please. Yeah, so I'm Alex. I'm a cloud chief technologist
within our cloud practice and look after our go-to-market strategy and portfolio around public and private cloud. And I'm Martyn Bullerwell, officially vice president of data and AI for Telefonica Tech, UK and I. I look after our data and AI practice within the UK helping our customers drive real value out of data and AI programs and projects. Formerly, I was a founder of a company called Adatus that was a niche specific company in data analytics and data engineering. And we've been through that process where Telefónica required us. And we're now heavily embedded within Telefónica Tech, helping drive even
more value to a broader range of customers through the Telefónica Tech ecosystem. Amazing. And we have a third guest, but before we introduce her, I'll make everybody wait a little bit. I also want to highlight that for the first time in this podcast, we use the co-pilot researcher agent to learn more, not only about Telefonica as preparation for this conversation and Telefonica tech and that transformation, but also about our guests here. So Alex, you said very well,
you're the cloud chief technologist. Here it said also that you're helping. Telefonica Tech become an industry leader in cloud optimization and FinOps. And you also won an award last year as the MSP Partner of the Year, validating your leadership driving cloud cost efficiency and innovation. And you often emphasize that reinvesting these savings into new cloud innovations for customers is the key. Is that accurate or not? It's all true, absolutely. We've been really driving FinOps quite hard since I joined. So I've only been
here sort of 18 months, didn't found a company, unfortunately, that got acquired. But yeah, I joined and really sort of started to focus on some of those enterprise use cases and driving sort of FinOps optimization. yeah, we got to a good place with IBM. We were the first, I think, MSP customer for them in terms of signing up to the new program and then bringing that to a lot of our CSP customers and also sort of midsize customers as well that want to get into FinOps.
Yeah, it's been a real success and we picked up the Partner of the Year Award and that's part of it. All true, definitely. Congratulations. Thank you. Co-pilot win, we'll take it. Yeah. Let's see if it goes two for two here. So Martyn, it says that your leadership has led Telefonica Tech to attain the Microsoft Analytics specialization, demonstrating extensive expertise in designing and managing analytic solutions, which for your title of Vice President of Data and AI, I think that's right around your wheelhouse. It also says that you're a thought leader in AI and you shared recently insights on aligning AI projects with business goals and highlighted the importance of data quality and governance for successful AI outcomes. How does that sound? mean, this is fundamentally why I love AI because it makes everything sound so awesome, should we say. But yeah, mean,
all those things are factually correct. If you ask an Englishman to say what they've done and how they've achieved it. I'm sure they'd be a lot more modest than that. But this is why we use CodePy. This is why we use AI to help us upsell. But no, it's all factually correct, yeah. Wonderful. Well,
we have actually a question for you about your use of AI. But before we head there, as I hinted before, we also have a very special guest. For the first time in our podcast, we're featuring a partner development manager, which are really a key strategic pillar of building those relationships with our partners. have Tijana Ellis with us today, also in the UK. Tijana, how are you doing? Do you want to tell us a little bit about you and your role as a Partner Development Manager, also known as PDM? Thanks for having me, Fede and Adam, and great to be here with you today. So yeah, my role as a PDM and Partner Development Manager is
all about working very closely with partners to drive joint success and deliver for customers. And the way we do that, I've been working with Telephone Architect for close to three years now. So we basically align our strategies, we co-create solutions, we work on co-selling with our teams and unlock new opportunities. And so my role is really all about building these strong, trusted
relationships and helping partners grow through mutual collaboration and innovation. That's exciting. So Fede alluded to it earlier, but we've really been working to integrate Co-Pilot both into the preparation of the episodes and then into the conversations as well. We'd love to hear from all of you, how you're kind of integrating Co-Pilot and AI into your daily lives. Maybe Tijana, you want to go first, then we can hand it over to Alex and Martyn.
So from my perspective, I use Co-Pilot every day and what I find with it is the more I use it, the more I find, keep finding new ways how it can help me. Whether that's catching up on meetings, using it actually to craft the documents, I use that a lot. And also for searching things and connecting the dots, I find it absolutely saves me tons of time just in everyday work. Obviously that's very much related to the nature of my role,
but I'm sure Martyn Alex may have slightly different ways of how they use it. Yeah, I guess I'll go next. So my journey with generative AI. So I'd like to split our use of generative AI and what I call traditional AI out. Because obviously generative AI is the cause of all the hype of recent years, which is obviously copilot is very much part of that generative AI piece and the traditional AI. But I guess my first interaction with generative AI was, I think I was on holiday. My brother actually runs a holiday villa company over in Cyprus. He was saying he was struggling
with marketing and actually wasn't very good at writing marketing content. None of his team were very good at writing marketing content. ChatGPT had been released and we said, well, let's have a go with it. We're looking at work. It's quite interesting. Let's see what we can use it for. What we started doing was asking him to write advertisements and marketing material for his website that would sound exciting, help sell villas or rent villas with hot tubs. And actually,
I think we probably spent 12 hours and 12 shots later, a good kind of time trying to look at it and have a really good laugh with it. And actually, what we found was it was impressive, really impressive for that kind of content generation. And, you know, that was my first dip of my toe. And what we did ask it to do, was write some of these adverts in Cockney rhyming slang. I'm not sure if you guys are familiar with London Cockney rhyming slang, but chat GPT is, it's, you know, if you've ever watched any of the films like Snatch or those kinds of films around about London gangster type films, they use this kind of slang word. And it wrote these adverts in Cockney rhyming slang. And I was like, this is going to be a fascinating game changer for us and marketing and how it's all going to work.
And that was our first kind of dipping the toe into the water. But since then, internally for us, Generative AI we use, we've built our own Generative AI solutions that help us do bid management, bid processing, help us write responses for bids. Our answer to AI at the moment is just roll it out, start using it, start using Copilot. It's going to become the next word. And again, all of you are probably a bit younger than me, but when I first started my career in IT, word processors were kind of going out and Word Perfect and Microsoft Word was being released and they were like, go on training courses, do we need to adopt this new technology? It feels like that with Copilot, everyone is going to be using it. So let's just hurry up and get everyone adopted on that so we can start doing the real game changing stuff. So yeah, my whole team within Daytrain AI use Copilot. They use the GitHub Copilots to help write code.
They use the Copilot in Outland to help manage their lives. We use the, we use it a lot in teams and summarizing of documents and creating presentations. So we use it all over the business very successfully. And as I said, kind of most customers we get an opportunity to speak to is look, don't delay, get on it and you're going to be behind if you don't start using these technologies. So yeah, we're using it a lot. That's exciting. That's exciting. Yeah. I guess just to sort of echo, I guess a lot of that, yeah, it's become part of everyone's sort of day-to-day life. I mean, copilot for me, especially being a more of a technical person, you know, my brain when it comes to being creative, isn't always as good. So having a tool like that,
that can just quickly pull something together, review something, maybe change the format in condensed words when we're trying to get responses and bids out the door. And not so much, you know, come up with the content itself, but just give us the outline, the framework of what we want to use for that piece of content and give us some ideas and stimulus and then review it with Copilot, that for me is awesome. I know there's a lot more we can do to it, but when we're looking at trying to build solutions and explain technical concepts or even just to get an explanation of how something works from Copilot to start with, you can get it in a format that works for you and interrogate it and question it. And it just saves so much time where you'd have to do all that yourself.
So that's been the real game changer, the speed that you can access information and we can go back to the old days of tech net and things like that, but we had CDs and things that we were loading. Now you can just sit and have a conversation about a technical subject and yeah, love it. Yeah, it's fascinating. I would agree with you, Alex. I find myself all the time where the speed at which you can get to an 80 % solution with Co-Pilot or ID8 and kind of like get your brain unlocked instead of just staring at the screen and saying, where do I start is amazing. yeah, thank you all for sharing that. That's always exciting to hear kind of how this journey continues as we all sort of watch it evolve in real time. Yeah. And watch ourselves evolve with it as Martyn was saying, you know, comparing to the adoption of Word. Now going from the individuals to the organization. And I think our researcher put it in the same way that Tijana put it when we were preparing for this conversation, which was Telefonica's evolution from Telco to Techco. Can you tell us a bit more about that story?
And the why behind it, why do you evolve that mission? Yeah, I'll put that up. So the history of Telefonica Tech is still quite a young company overall. I it was carved out of Telefonica to really just provide that technology services back into the organization. And within Spain, obviously with the dominance that Telefonica has there in that market, that's quite a big operation. So Telefonica Tech became like this sort of huge tech services company sort of overnight and that's continued to evolve and grow. And then the expansions that have then happened through acquisitions as Martyn pointed to as well, building out the UK capability has then given us these five practices that we have in the UK. So around cloud, data and AI,
we business applications, which is another acquisition around dynamics and power app. We have a strong cyber security practice as well that leverages a lot of the Spain market and the big cyber business that Telefonica has there. And then digital workplace as well, we're looking at Copilot and Office 365 and we have a quite a heavy sort of Microsoft slant, but also been able to build quite a strong presence off the back of those acquisitions that came with things like a strong data center and infrastructure background as well on managed services. So we've kind of evolved from that state, I guess, and integrated those acquisitions and over the last sort of few years really built up really what our story is. And that has really focused on solving complex business problems, which really focuses around bringing all of these elements together. So very much how Microsoft operates and being a bit ex-Microsoft myself,
understanding the landscape of, here's all the tools, here's all the products. And now for us as partners, let's build something out of that that addresses a customer outcome. And that's where we are and our new sort of wave of portfolios and offerings and with our new practices that have come online over the last year, that's really starting to accelerate and grow. It's an exciting time for Telefonica and Telefonica Tech. There's a lot happening. There's a lot of globalization as well and understanding how things that are being delivered in Brazil and Peru and Chile can be then leveraged and adapted into the UK as well makes it really interesting. So for me, we're like a mini-GSI,
but with a lot of power. Anything you want to add to that, Moin? Yeah, I think I was going to pick up on the mini-GSI piece because, you know, being acquired, I guess, from a founder owner of a relatively small niche, I'm to call ourselves a Be-Like Consultancy for those that remember business intelligence. You know, one of my main concerns for my team and for me was, you know, we don't want to be swallowed by a large corporate that are just kind of come in and tell us how it needs to run and kind of become dominating. And I think what's really nice about what we've done, certainly in the UK, is we've kept those five practices.
And we've kept the nimble agility that we can't with a smaller kind of organization. But we've got all the backing of Telefonica Tech in Spain and Telefonica as a kind of global company. So we kind of think big, but still act small, if that makes sense. So we still got the agility that helps really us drive value through our customers, be innovative and pivot when things aren't where they should be and we've always, from a data AI perspective, we've built the whole business for the last 19 years based on Microsoft technology. And it's always been great to us. it's even better that we can continue and grow beyond the reach that we had before, being able to help drive data AI solutions into a number of other organizations we may not have access to prior to being part of this bigger group, while still maintaining the culture and having fun doing it.
That's great. It's certainly impressive just to see the breadth of offerings that you all bring to your customers, really kind of across, I you mentioned your five practice areas and having kind of that robust offering. I think is really is key to help your customers drive that modernization that we know they're all seeking. One of the specific questions we'd love to learn a little bit more about is kind of what drove your decision to move your customers from your data center and into Azure and then beyond the migration. If you look at sort of the both modernization and incremental value add capabilities that you've been able to either offer your customers or maybe that your customers were coming to you and saying, hey, I need this or I want this.
Would love to like learn a little bit more about that journey and then really kind of some of the key outcomes and benefits that you were able to deliver for your customers as a result of that. I think from a cloud migration perspective, and certainly we look at the DCO program, which has been a real advantage for us and has been really helpful in driving that and being able to finance some of that as well in terms of giving us an incentive and a motivation to really push customers towards cloud. That's been great, but I think the customer at the end of the day drives that strategy. And one of the things that… we've seen is, I guess,
the days of those big, huge migrations maybe has gone and it's now becoming a much more of a workload conversation. Customers are looking for the right place to put workloads that suit them. And I suppose our job at the moment is to give them all those options. And that's where our philosophy is. certainly from my point of view, workload freedom is a key part of our strategy and our message is to say, look, where's the best place to put this workload and let's put it there.
but we want to make sure that customers have access and understand how they can build those sort of hybrid multi-cloud strategies and get the best out of Azure and say, right, I want them to have a cloud option to me. I want to look at these cloud technologies. I want to look at the innovation I can do with cloud native. So help me get started in that space. Help me look at this and then we can help them move workloads
from on-premise data centers into Azure and then start leveraging some of the data. I had capabilities that are available there as well as getting them started on DevOps and automation and just trying to optimize a lot of what we do. As I say, the programs and things that we have, and that's not just DCO from a partner perspective, but access to the Azure migration funding and Azure Innovate funding are key levers to help customers explore that world. And that's what really drives the acceleration of those moves. We are still seeing those migrations happen and we've got large customers that are sat within our data centers that we host critical workloads for, especially around the NHS. And now we're starting to see them pivot some of those workloads to Azure and start to leverage some of the capabilities within Azure to be able to innovate and modernize and also optimize those costs. And we mentioned FinOps, that's obviously a big part of it, but we also applied a lot of
security around that as well given that platform then to go and explore and modernize. yeah, I don't want to say that we're really sort of driving customers out of data centers because if they want to stay, they will. There's been some interesting developments in the market that may accelerate some of those decisions around cost and price increases on certain hypervisors. But that's what's so great about us as a partner having these options, being able to present all these different capabilities to customers, give them that freedom to pick the right place, give them the incentive funding to say, right, let's go and start up that assessment and do that migration, and then reap the benefits of that overall in terms of reducing the cost of operation of their workloads. And I think from a data and AI perspective, we were a very early embracer,
should we call it, of Azure. And actually, a lot of that was if we roll back to the time where we were talking about Azure Data Warehouse when it first came out and prior to that, was Azure, not Azure, there was FastTrack on-prem and there was stuff you'd have to buy to manage huge data warehouses and large volumes of data transactions. And it was huge upfront costs for customers. So actually from a data platform and BI analytics perspective, Azure was a very simple scalable answer to helping customers move large volumes of data with our huge great technical investment in on-premise architecture that they knew would grow. Obviously, we used the term limitless scale at the beginning. I'm not quite sure that's a fair term to use because I'm not sure anything's limitless, but I think we've got to be realistic and customers really very quickly and even some of the customers I like to talk about at the NHS realize that when they've got huge volumes of data, and they need to spend too much capex to scale something that will be fit for purpose for the next five years, the cloud solution made a huge kind of obvious step for Azure. And to kind of bring that to life and I'll roll back a huge amount of years, I'm trying to think when we did this piece of work, but we worked for a large music label, I won't say who it is, in fact, I don't know any longer in business sadly, but they ingested data from Spotify, from iTunes, from all the streamers, everyone remembers Deezer and all of those seven digital, there was loads of them that came out. So every stream of any music was owned by this music label who happened to own
the Beatles, so huge amounts of streaming. We ingested that data and we were trying to analyze that data. We were looking at types of customers that this label had. We helped use that data for marketing. And there was billions and billions of rows of data, as you can imagine, this is worldwide data from every streaming device in every one's home office location.
And what we found was if we had to reload one of the major tables, the FAT tables, on the single kind of processing kit we had on-prem, it was taking up to seven to eight days to reload this FAT table. Whereas when we first started using Azure Data Warehouse, we put it into Azure and said, let's see how quick we can scale out using a parallel threads. We could load that thing that took seven days in about two or three hours which was game changing for them, even though it cost them a lot of money, it means they could actually test and answer the questions much, much quicker. So, you know, for us, was, was from that point onwards, it is a no brainer, lower cost of entry for all our customers is scalable. We don't have to worry,
you know, gone were the days of having to go, how much sand storage do we have to buy to take the warehouse for the next six years? You know, that's a, that's a dim and distant memory for the data and AI projects, thank God. So really driving driving a lot of value from day one. So unlike Alex, where we have a lot of customers that have to think, well, you which one do we use and stuff, actually just the benefits to scalability and performance outweigh almost every other, every other problem you may have from a data platform perspective. And then when you move that to AI, when you think about the amount of processing and GPUs and things you'd need to run that on-prem, most organizations, unless they've got reasons to be have things processed in the edge or to have things processed very, very securely, again, would you want to invest in that kit at the moment with it being new and work out what you need to do? So again, for us in the day-to-day practice, Azure is a very straightforward answer to a problem. Yeah, yeah. I think that's really well said. think especially too, and that's a benefit
we hear from a lot of partners that have gone through this transformation is really just the new tools and capabilities that are at your fingertips with really no or very limited development costs, right? In the sense that the solution's built and it's more about how do you adopt that and then bring that value to your customers. So I love that story. Thanks for sharing that, Martyn. We did have another customer actually, sorry, I could talk for hours on this, but we did have another customer who wanted to do everything on-prem, this is again going back a little bit of time. SQL Server, SSIS, all of the kind of old data platform stuff that you put in. And we could, we said, what we'll do is we'll turn it on and stick it into Azure, we'll scale it up in Azure, and we'll wait for you to get your hardware ready. We got to six months and they still hadn't managed
to rack their hardware and their data centers. And we developed the program, finished the project and put it into production because we couldn't wait for them to rack the kit that they needed and the storage they needed. So obviously they sat that off and realized that Azure was the way forward. And since then, that customer has moved loads of their workloads across to Azure because they've realized that actually just the speed of being able to get stuff up and the lack of having to do capacity planning is super important. Yeah, yeah, it just takes that whole headache off the table. Yeah, that's exciting. Tijana, I'm curious from your perspective too as a PDM,
how you've kind of seen the relationship evolve with Telefonica as they've made these investments and these incredible transitions. Yeah, I mean, what I was going to jump in there, so thank you for following up, is I mean, you have heard the way Telefonica Tech is set up and just the sheer breadth of capability and depth of capability as well is very much aligned to Microsoft. So, you know, even the type of practices they have is very much aligned to how Microsoft is set up. So it's been quite easy to just have the alignment of strategy as well, because know, the telecommunication has been very, very intentional and focused on where to build a capability, how to be, you know, how to acquire the deep boutique capability like Datis that Mati was a part of. So… the capability was there, but what has also been really, really great, you know, in the last couple of years since we've been working together is just that speed and agility at which the phonic attack have reacted to the, you know, to the technology advancement, but also from our perspective, like the programs, like DCO and the others that got mentioned earlier, like being very quick to adopt those, to make usage of them, to embed them in their own, you know, go to market and how to give the most value to the customers.
So there've been like that alignment of strategy has been absolutely critical naturally, but just the way the speed of execution, like on the operational level has been absolutely key as well. And for me, having that strong AI data and AI capability, especially in the last, I would say probably 12 to 18 months, has been absolutely critical because the customers want to hear, want to understand what can AI do for them. The reality is their maturity of adopting AI there and then is not, you know, it's on a full scale. Some customers are really mature, some are not quite yet there. And the good thing is, Terraformica could just bring them on the full journey. It didn't matter almost where they were. So, you know, if the customer wants to use AI and their own friends, they can take them on the whole journey, migrate them into Azure and then help them adopt workloads. If the customer is already in Azure.
And then they want AI because they don't have it yet. Again, guys can help them with it. So just that flexibility in terms of how deep do you need to go with the customer on their journey has been just, I mean, it's been fantastic to work with them as as a PDF. And the other thing is what like the ratio goes both ways, right? What's, what has been really helpful for us is to just hear from the ground, what do the customers need? So we can, you know, we can feed that back internally to our engineering and product teams. You know, we have to find the challenges, you why don't you have this and, you know, customers are asking for that. And can we have this selling be quicker to market? So for me, it's been true partnership, you know, that has worked both ways. Thank you for sharing that. And
I forgot to mention in the beginning, this whole conversation is happening because of you, Tijana, because you reached out and said, I got a great story to tell. And so far in the few minutes that we spent together, I can see very well what you said, the great alignment that we have, the strategic pillars that are very well aligned between the two companies and how both of us are benefiting from that. So really appreciate you making this happen. And going back to the stories that you all were sharing, I actually want to pivot a little bit because I know that our listener may be saying, yeah, you're just throwing the highlight reel, right? Like the best customers, the best stories and the most biggest achievements that you've all had. But
if we were to... switch that a little bit as the customer or the listener is thinking about how to apply this with their customers, right? They might be thinking, okay, this might not happen because of this, or this is the biggest roadblock that we have. So tell us a bit more about those challenges or those roadblocks that you've faced and how have you overcome that as Telefónica Tech? I think it is, there are definitely challenges and I think we've all experienced that.
Customers do engage with, have their own preferences and their own sort understanding of what they do. I think, you know, you look at cloud, as Martyn said, some of the stories that we've seen, you know, there's clear benefits of moving the cloud and leveraging cloud, but you've got to be ready and set up to be able to do that, to take advantage of it. If you haven't gone through that, that thinking time, built a strategy, really started to understand about what you're looking to achieve, you can hit some real problems. And certainly customers that have gone in maybe a bit too quick or just try to migrate a bunch of workloads and then suddenly seeing costs getting out of control is something that's happened quite a lot. obviously, we can help with those situations with those spin ops type tools and solutions to get a handle on costs because it is a challenge.
But also just having that confidence that actually you're not going to get yourself locked into a situation where you don't have the ability to move or bring things back or you're taking a step too far that puts your company at risk should there be problems. So our job is to really help customers navigate that wider challenge and do that evaluation that they do. If you've worked on the customer side, you have to look around, you have to try and understand and make a judgment on where the best place is to put your key applications. And that's where the partnership really comes to bear because it isn't just then us coming in, we go in alongside Microsoft and present the whole Azure story and what it can do and help to define solution architectures and patterns that can really address their need, whether that's specific to a sector or to a particular use case that we're working on. That's where we can kind of really work and that sort of three-party relationship with the customer comes to bear. that's sort of where things are at the moment. It is a challenging market.
There's a lot of pressure on cost and customers, I think, looking for some real sort of certainty and flexibility and portability in that. And the more those solutions that come out that enable that ability to not get stuck and not get locked in or get caught out, maybe that's holding back a little bit of the acceleration of workloads. And there's quite a bit of governance and bureaucracy around of cloud at the moment. But we try and take a workload approach to it so we can unlock that and say, right, this workload makes a lot of sense. There's a lot of tools out there we can take and modernize the application, whether that's through containerization or cloud native capabilities, we can show the cost and give that confidence to then make that jump into cloud and drive the innovation. And that's the challenge I
think for all of us at the moment is to keep that momentum going and keep that innovation flowing and not get caught out that sort of fear of not knowing whether to make a jump or not or take a big commitment. We drive that flexibility approach and that's working. It's not easy at the moment, it's a difficult climate. And I think from the AI perspective, there's some real challenges around the hype. Everyone is talking about it. I did a presentation at our UK kickoff around, I don't know if anyone remembers Generation Game, in fact, you probably won't because it's a very British program. But fundamentally, it's a program where you put up a number, they used to have to remember a number of different items, and every item they'd remember, they'd get with a surprise.
So I put all of those up, these pictures of things up and there was a bed, some roller skates, a washing machine, a fridge freezer, a teddy bear, mattress. There was all sorts of random things that you wouldn't go. And the thing that they all had in common was they were all powered by AI or AI enabled. And you're like, how am I roller skates AI enabled? I really, just don't need, I just need the wheels spin. That's it. I don't need anything else. you know, this is about when they stop. Well, yeah, and maybe what resistance they have and maybe what maximum speed you're in, all sorts of stuff. But I think that this has been a real challenge for us and businesses is
that you've got people who going, you've got to adopt AI, you've got to adopt AI. And we've got people who have come into us, customers that come to us and go, we need to adopt AI. So I had a conversation at a CDO summit recently with a customer who basically said, I've been given the role of chief AI officer, and I've been asked to implement AI in the business by the CEO but I don't know what I can roll out copilot, which might tick that box, but he wants like game changing AI without any use cases, without any understanding of what data you want to have, what problems you need to solve. There's some real challenges and there's some real challenges around
going to customers and doing AI proof of concepts and pilots and programs that should drive value. But they look at the return on investment and actually they're just not there because running AI is expensive especially if you're building large complex things. Or we've got a number of customers that have come to us and they've come to us with specific use cases, but they just don't have the data and the quality of data to back that up. And I think within the AI landscape, this is one of the biggest challenges. The co-pilots are absolutely brilliant and they get adopted and they should just be adopted and we need to roll out and train people and get them to adopt those. But what I call the game changing AI, the stuff that's unique to us in each business that needs some thought, needs that quality data isn't just plug and play. It is, let's actually
think about what we're to build, what we're going to achieve, how much it's going to cost to run. And we can't be looking at it in a big challenge. You can't be looking at it, in my opinion, like a way to trim down people. It's not a way, should not be replacing people, in my opinion. It should be augmenting and improving productivity and improving what we do and the outcomes of what we do, but not it shouldn't be looked at as a cost reduction. And that's a challenge. That's a real challenge. Yeah, I completely agree. And I think the one thing I'd add to that,
because it's a conversation we're having with lot of partners right now and customers, is the fact that you start with copilot, right? And you essentially look to push copilot as far as you can go from an agentic standpoint and say, what can I get out of this? And I think the thing we are seeing is it's so powerful, just like that researcher capability that Fede mentioned it can do so much for you before you have to take that next step into more enhanced AI services. I think it's great to hear you mention that and validate that because that's really how we are seeing our customers and partners maximize their success. Yeah, absolutely. You see headlines all the time, but I saw a headline yesterday, the day before 80 % of AI projects are failing. And I think that's because no one's setting them up for success. The real big push is and we've had a couple of POCs we've put out that actually would have driven real value, but haven't gone through the right governance, haven't gone the right process before we get there. One, we couldn't roll out because of ethical usage of data. We had a really good
use case to actually help individuals not drop out of courses, educational courses. Really powerful use case, but ethically, were we using the data in the right way? They hadn't approved and signed up for the data to be used in that right way and also who was going to go and have those conversations. So they hadn't thought about the whole process. I think governance, when you're trying to roll out things beyond copilot, mean, copilot still needs a lot of adoption and governance. There's a lot of security risks around it, making sure that you're not exposing it to the wrong people. had examples, sorry I'm babbling, but we had examples of when copilot first came out, we had employees, I won't mention any names, but some employees that were saying, oh, we use it for put the code in to help us write our code. We're like,
no, no, no, no, you can't do that. You need to be on an internal Azure one. So there's all sorts of interesting challenges with AI. It's opening a whole new world of arguably productivity, but pain as well of security and management of data and quality of data and governance and ownership and all of the stuff that I frankly have been banging on for 18 years, but no one's been listening. So yeah, we're finally there. We're finally talking about data quality. We're finally using Purview to...to drive customers and it's a conversation we have all the time.
It's preparing them for kind of, I'm going to keep going game changing AI. But you know, the stuff that's really going to make the difference in my opinion, the stuff that you can uniquely identify yourselves with as a client. And along what you were saying, Martyn, in terms of security and compliance, can you tell us a bit more about what Telefónica Tech is bringing to the table? So we mentioned we have a cyber practice and that's definitely a key part of our cloud story as well. Everything we do has to be secure by design, which is the tagline that we attach to everything in cloud. But it's really fundamental, as Martyn mentioned, with new technologies, especially AI, where it's accessing a lot of data, those tools and those guardrails have to be in place. And
we're entering this new world of automatic code creation. And we're enabling a lot of developers that we'll suddenly start developing applications. I'm still waiting for that moment where we can really start building applications just from a prompt. And then they're going to go off and do all kinds of things. And we need to have that security in place. So building secure landing zones, following the CAF, using the enterprise scale patterns becomes really important. And then as we're starting to optimize customers, there's a lot of drive to again, to save money by using automation, by getting self-service in place to customers that want to spin up sandbox environments to do all kinds of data research. we have customers that do those
sort of trusted research environments and they have to be really guarded well. And what we try to do is leverage those patterns and technologies and bring that to the wider enterprises as well. So we're taking best practices from specific niche use cases and then trying to feed that in as part of our core messaging around cloud adoption. It's certainly challenging. I think
it's been challenging for a while, but now we've got new vectors that we need to be aware of and make sure that we're taking that holistic view. We're feeding that back into the strategy as well. Strategy, no matter how much we talk about cloud strategy, it won't go away. You still need to be doing it. And it's something that you think would have evolved over the years, but just having a clear view of your cloud security policy, setting yourself up with a strong foundation is still the key to getting it right. And that's what we try and help customers do. And I think that's the story of being kind of stronger together. The kind of data and AI
practice, we've always very much focused on the business outcomes and the value that we can drive from data. And with Alex's team in cloud and with the cyber team, we can really make sure that… the customer environments are secure and set up as I said by secure by design. And then the data and AI team can be safe in the knowledge that what we then build is securing it internally and helping make sure the customers and their employees are well-trained, well-managed and look after the kind of platforms and the data that we put into there. And I think from a governance perspective, that has been a challenge I think from a data platform perspective for the 18 years I've been doing this, or 20, sorry, 22 years I've been doing this. So, I think Microsoft have recently bought out Perview, which I've talked, I think I touched on earlier, and that's really helping bring data governance to the forefront. And what AI is doing is making customers really realize that the data will be available to the masses, their business data will be available to the masses within their business, and they need to ensure data is accurate and secure and trustworthy.
And that's leaning really well back into the story with Purview to say, actually, we need to put Purview and we need to put platform and we need to make sure your data is secured through a kind of sensible design. But we all know what's in it and we have data owners. So what we've been doing recently with a lot of our customers is starting to talk about kind of data culture and data products which is kind of an evolution of the data mesh, if anyone's familiar with data mesh design. And it's allowing businesses within the defined security principles that Alex was talking about earlier within cloud and the governance we can put in place to actually start building their own data solutions, own reports, their own analytics and their own AI output. But it does take time and effort to get that quality governance in place for customers. And what's really nice is that AI is
bringing that to the forefront and making us all very aware of how important security governance is. I'll just add as well, just to that, one of things that I really liked what the Alfonico team did, and I know some of the partners who had a similar approach, security is embedded and it's almost like the customers have to opt out if they don't want certain security to be embedded by default. And it just it's a small tweak, but it just gets the customer, the customers really think like, why wouldn't I want this? And if I don't want this particular security solution, do I have something alternative? It just needs to be there every step of the way, because we know from our perspective how paramount security is maybe now more so than ever. It's just, really like that
slight change in approach that Telefónica embedded a little while ago, but because it just sets the bar with customers where they really need to be. And if they don't want to be, they kind of need to really think hard, why not? And what is going to suffer if they don't go, you know, if they don't really secure their environment, whatever it is, whichever aspect of it it is. Yeah. Yeah, no, absolutely. I think we all agree like security is just paramount now. It's table stakes.
You have to have it and we're continuing to invest heavily in that. And we see our partners that make security practices, you know, a core component of their customer offerings I would say are frequently driving the highest growth. So it's great to hear you all are focused on that area as well. So maybe one final question here, because I know we're about at time. How do you see the role of hybrid cloud evolving? And maybe as part of that even adaptive cloud where we are really working to build a world where our customers can leverage solutions across multiple environments, hybrid, multi-cloud, Azure.
How do you see that evolving from your perspective and what are your customers, what signals are you getting from your customers there? Yeah, I think the hybrid multi-cloud is really the new norm because I think we've gotten from the cloud first mentality a few years back where the larger migrations were being done at scale. I think there's been a step back from that and a bit more of a mature approach to how you're setting up your cloud strategy and leveraging the right environment. And that creates a heavy complexity for customers, which I think is why some of the, maybe the pace has dropped a little bit around just making sure that we're doing things right, making sure that we've got control of our costs and making sure that we understand what we're going to get out of these investments in cloud. Those decisions are taking a little bit more time. But absolutely, the hybrid multi-cloud is the way forward. It's just now sorting out what exactly do you need. I'm sure Martyn can talk a bit about those sort of AI outcomes and use cases.
And that's driving a new set of requirements into that thinking and into that strategy. So how the customers then sort of work out where we want to do that. And it may be that there's things still on-prem and in the data center and then there are things in cloud. And then how do I get some common
management around that, whether it's, you and tools like Azure Arc and different options that everyone's bringing out sort of help to bring that world together. And I think that common control plane and the cloud management platform, which I think a few years ago we probably thought might go away, but it's now sort of come back with a bit more vengeance because there's real value in now being able to manage multiple environments and multiple clouds and on-premise data centers through a single control plan and then optimize your support and management costs around that. Yeah, it's a new sort of world in a way. obviously with hypervisors being shook up and some of the ball-corm acquisition that's made some real noise around VMware has created a bit of a land of opportunity for suppliers and Microsoft within that as well. Look at AVS and that's becoming a real candidate for even smaller type of workloads that give customers more flexibility, more opportunity. So the options of a vast and I think that's the challenge for partners like
us is to really step up into this space and help translate what the hyperscalers and what the vendors are all saying and what they can offer and what they can do and then align that to the customer's requirements and their use case and their strategy and we kind of stitch the thing together into an overall approach and a solution and a service that really gives them what they want in as fast a time as possible. I think that's our challenge as much as anyone's. I guess from the AI and data perspective, mean, really most of our workloads are sat in Azure. And one thing we remember is that telling all of our customers that cloud isn't necessarily the cheap option. It may be the flexible option and it may be the option that gives you the full scalability, but it's not always going to be cheaper if you're willing to run your on-premise architecture for longer and sweat it a bit more but when it comes to the world of data and AI, typically stuff gets put in the cloud because of the scalability, because of the flexibility of it. But increasingly, we're having some interesting conversations around machine learning and where you may decide you have very point models that need to be very low latency. And you may need those kind of at the edge or on-premise near machinery.
We've got some manufacturing customers that want to have those kind of AI predictive models for predictive maintenance, predictive breakdowns closer to the machines where they need sub-second kind of response times. So we train the models in Azure and then we kind of push the output of that into maybe a kind of more local system. A lot of our customers are still kind of playing with that idea. But I think that's where the AI thing's going to take us is a lot more stuff at the edge, a lot more stuff kind of on your devices. You know, we've seen it with, better not pick that device up. But we've seen it with certain devices where,
you know, they're pushing more and more stuff within to your device. think I was chatting to someone from Microsoft yesterday, day before last week, around the Co-Pilot Plus PCs, where we're pushing a lot of the AI functions back down to the kind of device right in front of you, rather than you having to submit it to the cloud and it's come back. So I think that's all going to evolve over time. again, as Alex said, making sure that the right workloads, AI is an interesting one, at the right point in the right place to support a customer's needs which may be on their device, it may be on their own data center, or might be next to a particular hardware they've got.
So I think it's going to be fascinating how it all pans out. But yeah, it still definitely all in on this year. Well, thank you so much because you've shared not only about your personal and individual journeys along this Kyri Cloud, cloud and now AI transformation, but also the evolution of Telefónica Tech from Telco to Teco. And of course, thanks to Anna for bringing this story up to life and letting us share it with our community. And Martyn and Alex, really appreciate
you staying up late today in your UK time zone to chat with Adam and I. Thanks very much, guys, for the opportunity. We appreciate it. Yeah, thank you. Thank you all. Much appreciated. We hope you enjoyed today's episode. Stay tuned for more insights and stories
from Microsoft's data center optimization team. Be sure to stay connected at dco.microsoft.com or wherever you get your favorite podcasts.
2025-05-30 04:58
