AWS Investment in New Zealand’s Tech Future

Special thanks to organisations who support innovation and tech leadership in New Zealand by partnering with NZ Tech Podcast: One NZ HP Spark NZ 2degrees Gorilla Technology

Episode Transcript (computer-generated)

Paul Spain:
Hey, folks, greetings and welcome along to the New Zealand tech podcast. I’m your host, Paul Spain. Today we have the privilege of two people from AWS joining us, Amazon Web Services. First up is Tiffany Blomquist. Great to have you here, Tiffany.

Tiffany Bloomquist:
Thank you. Lovely to be here.

Paul Spain:
And also Rada Stanic from AWS in Sydney. How are you?

Rada Stanic:
Thanks so much, Paul. It’s great to be here with you and Tiffany today.

Paul Spain:
So before we get started, maybe we’ll start with you, Tiffany. Just a little overview of what you do and where you fit into Amazon Web services and this big, wide world of tech.

Tiffany Bloomquist:
Fabulous. So, first of all, Kai Ora, you can tell by my accent I’m not originally from New Zealand, but my boys and I have made this our home for the last two and a half years. I am currently the country manager for AWS and big part of why I’m here is obviously the incredible amount of investment that we have coming to New Zealand. From a AWS perspective, we believe our purpose is actually around helping Kiwis change the world. And that is through the investments we’re making in skills and infrastructure, a variety of things which I hope we’ll get a chance to talk about today.

Paul Spain:
That’s great. Well, that kind of lines up with the sort of things that we like to talk about on the New Zealand Tech Podcast. So that’s great. And Rada, tell us a little bit about yourself.

Rada Stanic:
So, I’m Rada. I’m a chief technologist at AWS, looking after our Australian New Zealand customers working on strategic projects that are of real great importance to our customers across different industries and of different sizes. I do have a passion for all things data analytics, generative AI and also modernization of workloads in cloud. And. Yeah, that’s me.

Paul Spain:
Fantastic. Before we jump in, of course, a big thank you to our show partners to One NZ, 2degrees, Spark, HP and Gorilla Technology. Well, let’s jump in. I think there’s really lots that we can talk about today. Definitely keen to delve into AI and where that’s going. And AWS has certainly a key role to play there. But also, for some folks, might not be the first name that comes to mind when they’re thinking of that, but when we talk hyperscale cloud, well, AWS is kind of the number one company and really you’ve set the standards globally. So I think a lot of listeners will be expecting me to ask you a few questions down that track.

Paul Spain:
What’s happening? When are we seeing your data centers launch and active within New Zealand? And what actually are you doing here within the New Zealand market? I think for a long period we really didn’t know whether New Zealand would ever end up with any of the hyperscale cloud providers coming to New Zealand. And then Microsoft jumped in and said, hey, we’re doing this. And AWS obviously made an announcement following that. So can we start there, Tiffany?

Tiffany Bloomquist:
Yes, there’s about seven questions in that. So I’ll do my best and let me know if I miss anything. So first of all, AWS is actually Amazon Web services. And we actually started with the concept that Amazon had been really good at creating a website to actually enable customers to buy product. Why don’t we actually share some of those skills with our customers? There’s a problem statement out there that other customers want to be doing this well. And so in 2007, AWS, which we fondly call it because it’s a lot shorter, was born and actually created cloud. So in other parts of the world, AWS is a well known brand when it comes to anything related to cloud in New Zealand, I think it’s newer that we’re actually able to be talking about the breadth and scale of services that we offer and the variety of different things that New Zealanders can actually use AWS for. I think the primary way people think about AWS is infrastructure.

Tiffany Bloomquist:
And what we mean by that is you don’t actually have to be managing the servers and patching and doing all of what we call undifferentiated heavy lifting that was required for you to operate a business. Historically, Amazon would do that for you and at a lower cost and in a more sustainable way. And that message around how we did that work and the benefits that customers were seeing, we have an unprecedented number of price reductions that we’ve actually provided to customers over the years. Most businesses can’t talk about that as part of their business model. That is where we realised there’s actually a lot of work we need to do in New Zealand to help from a brand perspective, really get that message out about what’s possible with Amazon and with cloud now, New Zealand in particular. One of my favorite things about it is, as I said in that purpose statement that we have about Kiwis changing the world is there’s incredible amounts of innovation that are in New Zealand that frankly we want to be a partner to help them export. And a big part of that is obviously creating the capability locally from an infrastructure perspective. So we did launch a seven and a half billion dollar commitment to New Zealand for a region to come.

Tiffany Bloomquist:
And I’d love to clarify what that actually is because a lot of companies talk about the level of investment they’re making. We’re very specific on what a region is and we don’t use those words unless we’re talking about it. And that is actually multiple different availability zones, minimum of three, which means there are different locations across Auckland that will have infrastructure in them that enables high resiliency and connective tissues across all of them so that customers can really trust in the experience that they have with us. This isn’t one data center being spun up. This is actually a level of investment that’s pretty significant. These projects are complicated to get done and they are a significant level of investment. But that is what we have committed to New Zealand and we are absolutely still committed to bringing that. I think one of the things that we’ve realised though, as a part of our journey to bring this infrastructure to New Zealand is as much infrastructure as you bring.

Tiffany Bloomquist:
If you don’t have the skills in New Zealand, people aren’t really going to be able to uptake a lot of those capabilities as fast as they can to really benefit their businesses. So we’ve taken a little bit of a different tack over the last, I don’t know, year and a half or so in looking at skills reports specific to New Zealand and talking about what actual employees and customers are saying about skills and what they need. And what we’ve realised is we also needed to make a pretty significant level of investment in skills. And we can talk about that a bit later as well in order to ensure that when we bring all of this fabulous infrastructure we’ve got skills for, it doesn’t mean we’re not still investing from an infrastructure perspective. We have a variety of different services that we’ve launched. You may be aware that we have a power agreement with mercury where we’re actually investing in their turatillim farms and actually sourcing all of our power directly from them, which is differentiated than just buying maybe energy credits because we’re helping them create the next new energy projects that will create even more power for New Zealand. And then last year we actually also launched our local zones, which is an ability for Kiwis to actually have access to single digit millisecond latency for very key workloads like 5G or some of these environments where you may want to have a little virtual gaming creation that we’re seeing a lot of customers interested in. So a lot is happening.

Tiffany Bloomquist:
A lot of investments are coming to New Zealand. We’re still committed to those investments and there have been quite a few that are happening along the way.

Paul Spain:
So talking to the sort of timeframes, what can you share on that front?

Tiffany Bloomquist:
Yeah, what I can tell you is that we are committed to bringing the infrastructure to New Zealand and that we are constantly looking at communicating around when the actual timeframes are. We’re careful around what we say about the timeframes, though, because what we want to ensure is that, one, the infrastructure here locally shouldn’t slow down any customers from adopting cloud. We have quite a few customers that are actually leveraging our Singapore and even our US west and mostly our Sydney data centers as well as Melbourne. So we have a variety of infrastructure already available for customers today, and we don’t want any of these new investments slowing down the amazing benefits they can already get from cloud. And then two, we know that it projects take a lot of planning and attention and time, and so we want to be really careful about when we actually communicate the dates that it will come. So hopefully we’ll be able to come out with more specificity around time frame, but it is still something we’re expecting to have in the next coming years.

Paul Spain:
Okay, was that a couple of years? Coming years, coming years. Okay. We’ve seen obviously some of the type to scale data centers sort of being built and so on over the last couple of years. Are you able to talk to infrastructure that’s appearing in Auckland and that’s being built right now, whether any of that is yours or whether that still hasn’t actually started yet, can you give some clarity?

Tiffany Bloomquist:
Because I’m New Zealand and Auckland in particular is a pretty small place. One of the things that we’re really careful about, though, from a privacy perspective and security perspective, is not disclosing the location of any of our data centers and not talking about the status of any of the construction of whatever those may look like. What we can do though, is we’ve arranged deep briefing sessions with particular customers and partners to help them understand what does a region actually look like, how do the availability zones work, what type of services come at launch? So we are doing briefings with customers to help them understand the more nuanced infrastructure architecture that comes together around that. But we’re not walking someone down the street and saying, look at that pretty thing. That’s part of the way that we actually ensure security.

Paul Spain:
Yeah, I guess there’s a level of that security by obscurity, because realistically, usually you can work out who owns what and where it sits. Right. And yeah, I think people think it was Auckland council that revealed resource consents for. Was it draining some wetlands or something or. You know, you can’t really kind of completely hide, can you?

Tiffany Bloomquist:
No, but we don’t want to hide. I mean, actually being able to have people interested in our region and talking about what’s coming and us continuing to commit to that. Coming to New Zealand is very important and we know it’s really important in terms of the level of investment. So this region should be continued to be talked about because it’s seven and a half billion. It’s supposed to create a thousand jobs for people. And so we take that commitment to New Zealand very seriously. We don’t make that lightly and that is something that we will be bringing to New Zealand.

Paul Spain:
And how does that break down? We sort of talk about a thousand jobs. A lot of this infrastructure is designed to run as autonomously as possible, but then there’s also a flow on from these sort of investments in other areas.

Tiffany Bloomquist:
Absolutely.

Paul Spain:
What does that look like in terms of those jobs?

Tiffany Bloomquist:
Yeah, so obviously there are jobs that are created simply to secure, maintain and run the data center for sure. In the lead up to construction, though, there’s a variety of different jobs that are created to actually do some of those activities as well. I would also say when we have a presence in New Zealand like we do today and we’re growing, we also can find additional ways to add value from a New Zealand perspective. So we actually have additional teams in New Zealand that are based here that are growing over time. One in particular is our engineering support teams. So these are individuals that actually provide support to us customers because of the time zone overlap. So we’ve got a growing contingent of employees that are actually supporting not only New Zealand customers but also us customers as well. We have some research and development teams that are based here.

Tiffany Bloomquist:
You may have heard about our partnership with Vector where they have been working on a software platform that helps to look at the way energy is actually managed across the grid. They have a product called Diverge from VTS is a subset of that. We’ve actually invested them and part of that was actually hiring engineers here on the ground to actually do some of that work. So the region is an interesting and compelling event and very exciting, especially for customers that care about regulatory preferences. But that announcement and that investment also means that other opportunities come up in working with customers and we can expand the number of jobs we’re committing to the region.

Paul Spain:
That’s great. And I guess looking at New Zealand, we’ve had some discussion, particularly over the last few years around the opportunity to build New Zealand’s tech sector, whether it’s attracting people from other parts of the world to come here and set up operations and so on, or whether it’s just having Kiwis work from New Zealand rather than travel off to Silicon Valley and other parts of the world. What do you see that opportunity being for New Zealand? You talked about R and D and engineers being based here. What are your thoughts on what that potential is? It sounds like you’re mostly talking at the moment. Maybe not big teams being established in New Zealand, but more sort of work from home engineers, which I guess could happen in any country. What do you see the sort of bigger potential for New Zealand and AWS?

Tiffany Bloomquist:
Yeah, for us it goes back to that purpose statement around helping Kiwis change the world. And what we’ve seen is that many Kiwis have innovative ways of looking at problem statements and they’ve got these fantastic ideas of what they want to see come to fruition. And that idea needs to be built out into a product. And one of the best ways that we can see New Zealand thrive is actually by getting that product to be a technology that can actually scale. Because then you’re not just selling to New Zealand, you actually can sell to the entire world. And as you see now, while we still have agriculture as the primary industry, the second industry is really tech. And that means that when we invest in the skills to Kiwis, we enable their great ideas to be built into software that can actually scale. One of the largest jobs that I have is actually coaching and mentoring and working with customers around.

Tiffany Bloomquist:
What does it mean to take your product and scale it far beyond the shores of New Zealand. It doesn’t mean we shouldn’t be investing in New Zealand 100%. But think what that level of investment, when you can scale your product to the world, brings back to Kiwis every single day, and then it creates jobs and then it creates communities and then it can create even a better brand for New Zealand on the global scale. So a big portion of my time is investing in those companies to grow those brands and help them sell overseas, but also grow their products around the world.

Paul Spain:
Rada, I’m curious for your perspective on this and this opportunity for New Zealand to grow our tech sector. And with you being based in Sydney, you may have a bit of a view on how’s New Zealand doing. If we were to put ourselves sort of competing with Australia, we like to say that we punch above our weight in varying areas as a country. But of course, Australia is a much bigger country, a bigger economy. And most of know if you look at the organizations, there’s often much larger organizations in Australia in varying ways, whether it’s tech or other sectors. How do you feel we’re doing as New Zealand, and what do you think we could be doing to compete better on that global stage?

Rada Stanic:
So I think New Zealand is doing really great. And it’s not just a statement, but there are data points and anecdotes and customer proof points that we have to support what I’m saying. Like Tiffany was saying earlier, we’re truly building this distributed cloud infrastructure with Sydney region, Melbourne region, Auckland region is coming up. There was a launch of Auckland local zone for those ultra sensitive workloads. We are providing customers with what we call customer experience in an on premises environment as well within their own data centers. So there is such a choice in terms of infrastructure. And I believe that New Zealand businesses are taking advantage of that and building. And know it’s interesting reading this exciting report around accelerating the AI conducted by our partner Access Partnership.

Rada Stanic:
We’ve seen actually that more than 90% businesses in New Zealand are, for example, looking to adopt generative AI in the next five years. And what I found really fascinating is that it’s also cross generational. So also close to 90% of workers, of employees are also looking to build in generative AI in their day to day jobs. And what is exciting, I think it’s also looking at Gen Z millennials, boomers. That figure doesn’t change very much. So everyone’s excited about the prospect of that. And you asked me to contrast that with Australians. There was one small stat that I saw in the report which actually says, for example, in New Zealand with boomers specifically, they are more keen to embrace and take generative AI in their day to day jobs than, for example, in Australia.

Rada Stanic:
And I do work with some of the New Zealand great businesses as well, and I do not see much of a difference. Innovative businesses want to reinvent themselves, they want to cost optimise so that they can innovate more. And I really see some exciting customer use cases emerging, which I’m sort of happy to talk about, whether it’s generative AI space or other areas. So I don’t see New Zealand as lagging at all, in fact, pushing the boundaries and innovating on many fronts. And there are data points in the report as well to support what saying.

Paul Spain:
Okay, that’s really encouraging to hear. We do like to keep up with our australian friends, shall we say? We definitely competition don’t like to be falling behind too far now. Sort of, I guess delving into this journey to the cloud, I’m keen to explore it a little bit more. And I guess I have a range of conversations with organizations at different levels. And for a hyperscale cloud provider to move into the New Zealand market, there generally has to be a fairly strong level of confidence that customers are going to be coming with you for that journey, because it is billions of dollars worth of investment. And the discussion I had a while ago, and I won’t mention any details, but it was an organization that had made some level of commitment to a hyperscale provider in New Zealand. Once they got into sort of the reality, and I guess this was letters of intent or whatever were signed at some point. Once they got into the reality of it and delved in, into the deep technical aspects of it, there was certainly some feelings that it was going to be a real challenge to maybe what they’d achieved, what they were initially looking for.

Paul Spain:
How hard is that journey for organizations that are thinking, hey, let’s get rid of our own data centers, or hosting with local providers and move everything to the hyperscale cloud. And I guess what are the realities that you’re seeing New Zealand, Australia and rest of the world? I don’t know which of you would like to speak to.

Tiffany Bloomquist:
Maybe I’ll start with the New Zealand point of view and have rada chime in. So I think one of the things that differentiates Amazon and AWS and the way we engage customers is that we don’t go through that whole process you just described of letters and intent and all of this public marketing. We have a very humble approach. That is, we need customers to be really clear on the value for them. And we have a variety of different ways that we can support them in that journey. And it’s much more important that we lean into where is the customer? What do they need? What are the challenges they face? Than we sign some massive data center exit proposal, because different customers need different things at different stages. So some customers, for example, absolutely have the engineering capability to literally leverage what I fondly call the Lego building blocks of AWS services. But there’s 200 plus services that are available to you and you have to understand the architecture, how to put them together.

Tiffany Bloomquist:
And you’ve got to have that capability. Or frankly, maybe you have the capability, but capacity in your team to do that work. And some of our customers embrace that, especially our software customers. They have those skills, they have those engineering teams. That’s a fantastic journey for them. Others actually need to be thinking about it a little bit differently. They need to bring in partners that have the capability to build what they need for their environment. And even others actually don’t want to build, they want to buy.

Tiffany Bloomquist:
That is really important for them. And so we actually have a capability called AWS Marketplace, where our customers, as part of their commitment to how much they’re looking to actually work with us, can buy other software solutions that are meeting their needs. You’re not locked in to like, you have to do this thing with us and it has to look this particular way. It also means though, that when we work with customers, we start with foundations, like literally cloud infrastructure. The best analogy is electricity. You should be able to flip the switch on and get the services that you need and flip it off when you don’t need them. You shouldn’t have your light running all the time, but a lot of customers have their lights running all the time. And for us, we want to figure out what are those workloads and scenarios where you can benefit from the scalability of cloud.

Tiffany Bloomquist:
And you should be thinking through what those are. We don’t want to just have you sign an all in. We want to think through how is this providing you the cost, the efficiency around costs that you need? How is this providing you with planning around the services to your customers? And also how are you able to respond to demand when all of a sudden we have game companies that their games go through the roof? How are you ensuring that you’re able for that gaming company to be able to respond to thousands of extra new users an hour if you’re not in cloud, they would have absolutely collapsed. So we’ve got so many examples of where the scalability, reliability of cloud meets a customer’s need. And we shouldn’t sign a whole bunch of things before we actually explore what the value is for them. The last piece that I would say for Kiwis, and then I’d love Rada to jump in as well. There’s an incredible desire to ensure that we’re doing things in a sustainable way and we’re thinking through the impact on the environment. We’ve got such a beautiful country here and all of us want to ensure that we are doing the right thing from that perspective.

Tiffany Bloomquist:
And yet, many companies haven’t actually looked at how sustainable it is for them to run their own infrastructure. And we know that most companies that move to us from an APAC perspective have an 80% savings when it comes to energy efficiency. Like massive amounts of thinking differently around the cost and the efficiencies of actually leveraging cloud. And so for us, we do see cloud as an inevitability, but we need it to be right sized for customers on their journey. And we need them to be aware of all of the benefits. Maybe not just financial, maybe not just from a tech perspective, but also from a sustainability perspective that they should be thinking about when they move to cloud, so that it is clear that they should be doing it. And they’re not signing anything in advance of doing some of that work. But we do feel that the majority of companies absolutely benefit from moving to cloud on a variety of those fronts.

Tiffany Bloomquist:
Rada, do you want to add anything to that, maybe with an Aussie perspective?

Rada Stanic:
Yeah, definitely. I think, you know, a couple of key things that Tiffany mentioned there. It’s like we always do what’s right thing for the customer, and that may be completely different things due to the nature of their business, how quickly they’re able to move, the skills they have. And as much as a technologist, I always am drawn to a technology aspect of something, a lot of it. That whole move to cloud is also influenced by the people and process side of things. So we not only provide all these services and building blocks and a platform that’s secure and scalable and resilient, but we also provide them with things like our cloud adoption framework, which guides them, depending on the levels that they’re at, how they can progress their journey to cloud. What I’ve observed globally, and in Australia specifically, is that those large enterprises, for example, that have multiple data centers, there are no questions anymore, should I move to the cloud and what are the benefits? But the asks that are coming our way is how fast can we move to exit those data centers? And then the conversation becomes, it’s actually fascinating to watch different sides of business, let’s say the companies that have legacy mainframe applications, and then they’re also building new and innovative capabilities around AI and data and so on. The conversations become, okay, how can we take those legacy applications that are running in your data centers and perhaps just move them to cloud infrastructure? So we’re not going to make major changes because it’s complex, we can’t find skills to unpack it all.

Rada Stanic:
So we’ll just move what we have with the smallest possible changes to cloud to start taking advantages of cloud. But then, on the other hand, with some of the newer applications and workloads that they have, let’s see, how can we build those to be cloud native and microservices based. So we have everything in between but the common thing is there really aren’t questions anymore as to why should I go to cloud, and security and resilience aspects of that have been answered. So one aspect, one end of the spectrum is those businesses with legacy workloads and applications, like I mentioned, mainframe. But what is also even more exciting is those businesses born in cloud. I mean, we touched on AI, and I’m surprised that we haven’t jumped into generative AI already because that’s the tone of every meeting and conversation these days. So companies like Leonardo AI, born in cloud, building generative AI applications that wouldn’t have been possible to be built without a scale of cloud. So building something where you can process four and a half million images a day using innovative technologies like AI and build an entire business model around that.

Rada Stanic:
So that’s sort of some of the trends that I am seeing, and I can tell you that I am answering questions of this nature. Is cloud secure? Is cloud resilient? Less and less. And the questions that are presented to me and my fellow Amazonians is how fast can we move? Let’s unpack this particular use case and workload and see how do we transition this in the most optimal way to the cloud. Those are the questions that I’m being asked, I guess. So that’s the perspective I can share from global and australian angles.

Paul Spain:
Yeah. Thank you. Now, you mentioned mainframes, and this is probably not something that most of us spend our time thinking about every day. What about the mainframes? But we do know that in banks and other large organizations, there are still these systems that they often haven’t figured out how to get them off their mainframes and onto more modern platforms. And I don’t know what that journey is when we can smash up all the old mainframes and so on, and there’s still that demand for cobol coders and programmers and so on. So some of this old stuff is still sticking around. So what can you tell us around? Is this mainframe stuff going away? And if it is sticking around, what can you do with it in terms of being able to put it into the cloud?

Rada Stanic:
So it has been proven large enterprises have moved their mainframe workloads to cloud. It’s been done a number of times. You mentioned shortage of perhaps COBOL programmers and expertise. And that’s why we have some of the partners that have built their businesses around helping companies move mainframe from on prem environment to cloud. So there are those specific niche skill sets, some of which we do have it in AWS, but we scale it through our partners. And I think Tiffany also mentioned a broad range of services and building blocks that we do have to help customers build almost anything they can imagine. We’re also focused on creating these managed services that help customers migrate from an on prem data center environment to cloud. So it’s been done in the airline industry, it’s been done in the financial services industry, so regulated industries as well.

Rada Stanic:
And it’s not a precedent, it’s a quite proven methodology. And again, the questions that I’m living through, the projects at the moment were, how fast can we move? And perhaps as a business, our customers are telling us we will not have time or we will not have all the money to just modernise that mainframe workload and turn it into a cloud native. But it’s more like, let’s just transition it to cloud so that we can reap the benefits of the cloud operating model, and then we’ll look over time how and when to modernise that. That’s a fair assessment of how mainframe workloads migrations have been approached.

Paul Spain:
Right. So you can kind of emulate the mainframe environments within your cloud for those who aren’t ready to actually rebuild, modernise things. Okay, there you go. I don’t have any workloads of that myself, but thank God it’s a tricky problem. It’s good to know. Although I did come across a firm fairly recently that had built all of their technology on a free version of Pascal language, which I found rather a curious approach.

Tiffany Bloomquist:
Yes, some risk in that.

Paul Spain:
That was a small business with quite an interesting approach.

Rada Stanic:
I haven’t heard of Pascal just to hide my age or so, but yeah, that’s awesome.

Paul Spain:
That’s so good to hear.

Tiffany Bloomquist:
Even better, I think one of the things, though, that I’d love to highlight about this conversation is I think historically, people may have thought about AWS as infrastructure. And one of the things that I think you’re hearing in this conversation that we’re changing is we actually want to have a really humble conversation with the customer about their biggest challenges. Sometimes that might be coding on a free version of software that they shouldn’t be, or other times it might be mainframe migrations. Whatever it is, the intention is that the customer is able to have a relationship with us where they are sharing the biggest challenge they have, and we are systematically going down. Now, for some software companies that is co selling into the United States, for some enterprise organizations, it’s the technology roadmap to simplify what they are doing, because it’s gotten so complex as they’ve grown over time. For others, it’s even how do we think differently about the skills we’re investing into organizations? Like how do we actually think about readying our team for things like AI and the future of what’s to come down the line? And so I think that’s why we’ve brought up this particular survey and it being so important for customers to engage in AI and think about it because the workers are ready for them. They see the future. I think by 2028 they said what 86% of workers feel like they’ll be using AI.

Tiffany Bloomquist:
It is the companies that need to actually lean in and invest. And it’s probably the number one question that I get from most ces and boards is how do I think about AI? Where do I get started? It’s so big. There’s so much going on. Everybody’s talking about it like bitcoin came along or other things have come along. Why is this one so incredibly important for us to get across right now? And I think it’s because it’s lit up in the eye of the consumer as something so useful in their day to day lives. It’s now become an imperative for businesses to get behind because practically people can see what it could change in their life day to day. That’s extremely exciting. But it’s also like kickstarting a whole wave of investment and technology and options, which is oftentimes very overwhelming for most companies.

Tiffany Bloomquist:
So one of the things that I wanted to make sure that we talked about in regards to AI and some of these big challenges that are coming is we’ve got a very specific process at Amazon that we always use when it comes to scenarios like this. And it’s always about working backwards from your customer. Now we use that to make sure we’re doing the right thing for the customer and the engagement. But I highly recommend, as a part of evaluating AI, you’re thinking actually, and it could be your employee as your customer, it could be your end customer as your customer. But how are you not just flipping on a switch and playing some additional licensing fees for stuff that you’re not even sure is going to work for your business? How are you actually thinking about, know, what part of the customer experience could I actually uplift? Or what part of my employee experience should I uplift? And if I do turn some of these things on, which could be the right thing to do. We’ve got amazing announcements from a variety of our AWS customers around AI, whether it be Seoul machines or Xero or others. They’re all announcing stuff right now. How do you ensure that you’re actually thinking through the way you’re going to measure that for your team so that you’re enabling them properly around it? You’re seeing the benefits to your customers or your employees and you’re getting a really good ROI on data usage.

Tiffany Bloomquist:
It’s how we approach a variety of different problems that we have. But I think in particular on this one, it’s all over the place and most customers are struggling with. Where do I start? I think it’s always starting with working backwards from the customer.

Paul Spain:
Yeah, I like that approach. Now, before we started, we had a little bit of a chat around your Amazon Connect call center technology. We know actually one of our show partners, One NZ, is using that. And this is an area that sort of fascinates me because I guess a lot of telecommunications companies and technology service providers have had varying challenges when it comes to looking after their customers, whether it’s making the customer feel as though they’re really important, when actually things are maybe outsourced to another part of the world. And so we’ve seen that change where more and more customer service is being brought back locally. But I’ve been really curious around all the possibilities that AI sort of brings in these situations. And I guess if you put yourself in that position as somebody sort of taking a phone call with a customer in a call center, it seems to me that there’s a lot that AI can potentially bring to the table to make those calls maybe a little bit easier, because the AI can trawl through past phone calls, past data conversations, all sorts of things, and sort of, in theory, hand them up on a silver platter to the individual that’s taking the call. And of course, it raises questions.

Paul Spain:
Should those calling in be talking to a human or talking to an AI? And so on? We’ve had all the terrible examples of calling up a call center and them asking, what do you want? And you tell them and they’re like, I don’t understand, tell me again. Or they forward you off to some menu that departments later. Yes, waste of time and so on. So, yeah, can you talk a little bit about that one?

Tiffany Bloomquist:
Yeah. So just like we were talking about AWS helping with undifferentiated heavy lifting of server management, I would love for people to think about AWS helping with AI as that removal of undifferentiated heavy lifting that most employees have today that AI will start to actually remove. Now, we’re in the early days of figuring out the best ways to do this from a cost perspective, because you can build a massive large language model and use the best of the technology, and it will cost you, but you can also scale it down and do it in a really efficient way. I love this example with One NZ because they’re deeply wanting to ensure, starting with their investment from an Amazon Connect perspective, that they’re uplifting their customer experience and they’re getting better and better at that over time. And they’re committed to that. And they’ve seen AI as a way to actually differentiate the engagement that they have. But they’ve actually realised there’s not one model to solve all of their concerns or problems. There’s not one solution that they can just slap on this.

Tiffany Bloomquist:
What they actually need to do is figure out how they really uplift those call center agents. So what they’ve done is they’ve actually created several models that they’re putting together to actually give the agent the most amount of information that they can in that experience with the customer. Now, some of that is looking at information from self help, publish, wikis and things to actually pull the most relevant information. Some of that is information about the account. But that means that the actual person talking to you on the other end of the line has a lot more information that they could really quickly looked up. And that speed of answering of questions has made a difference. In fact, they’ve seen a ten since. In just the last couple of months that we’ve put this together, they’ve seen a 10% increase in the amount of trust that the customer has with the agent, and a 10% increase in the perception of knowledge and information that they have.

Tiffany Bloomquist:
But it’s also a really great way to play with AI because you still have a person there evaluating what’s coming from the models, making sure there aren’t things like hallucinations, where they’ve put things together, maybe in a way that doesn’t make actual sense if you understand the context of the business. And they’re still able to have a remarkably different approach to the engagement with that customer and the quality of answers they’re providing them.

Paul Spain:
Yeah, I think it’s going to be fascinating to see how that evolves. And it’s been talked about in the past. Vodafone coming into one NZ have had some challenges around the perceptions of their customer service, which presumably is a reflection of shortcomings. So always interesting to see how technology can help on these fronts. Any other sort of New Zealand examples that we can?

Tiffany Bloomquist:
Well, it doesn’t have to just be large enterprises. I mean, I love the one NZ example because summer Collins and her team, we’ve talked in broad forums around what we’re doing on that front and they’re really leaning in and embracing and telling customers how they want to improve. But I think there’s also quite a few companies that are like in the startup space that are using it. You don’t have to be a large enterprise organization to do that. Clearhead is a great example of a startup that we have that is actually focused on mental health. Think after the impacts of COVID a lot of people have realised the importance of this in relationships. And so they’ve actually leveraged AI as part of their tool set so that they’re able to provide prompts to users around a variety of different content and assets that are available to them. And then they do also provide access to therapists from a worldwide network.

Tiffany Bloomquist:
But that’s again, a kiwi startup thinking about a way that they can embed AI into their business model that will enable them to scale and grow long term and obviously in an area that’s incredibly important as well. So I guess the message that I would say is there is not one size fits all for AI. People shouldn’t assume there’s one model or one way or one approach. In fact, the way that we think AI creates a whole set of options. Again, going back to that, we don’t think each customer needs the same thing. So whether it is a set of infrastructure, because you’re really well versed, let’s say you’re a data scientist, you need different needs. You have different needs and you need different tool sets. We have a whole bunch of chipsets and infrastructure that’s available to them.

Tiffany Bloomquist:
Let’s say you just want to turn on one of our partner models. That’s another layer. You can just engage with one of the models that we have today, like cloud, you have the ability to use that and actually take an instance from that partner of their model and put it in your own environment so that they keep their ip. But you use your data in your own environment, or there’s a variety of different ways that you can just lean in and use some of the different tools we have, like code whisper or Amazon Q. So there are a whole bunch of ways people can engage. I wouldn’t want it to feel like it’s ever just one size of business or one type of problem. We should be able to help in whatever way that customer needs.

Paul Spain:
You mentioned chipsets there. Yeah. Where do different chipsets fit?

Tiffany Bloomquist:
Yeah, I’ll start this one, and Rada can add probably a whole lot more cover. We find that based on the type of work you’re doing with AI, you actually have different needs in terms of compute. So we have chipsets like inferentia and tranium that are used for different things. If you’re actually training models and trying to tweak them over time, you need different types of compute power. And so Amazon has invested not only to ensure we’re keeping the cost of compute as low as possible, but we’re also differentiating different chipsets need for different types of work. So you have the best cost per the type of workload that you’re trying to move through into production. It just depends on what you actually need for the service you’re providing. Rada, do you want to add in on that?

Rada Stanic:
You’ve really articulated that very well. And when people talk about generative AI these days, the conversation very quickly just shifts to, oh, is there some fancy chatbot that I can use and ask some questions? And then I get these interesting answers. But there is so much that needs to go behind the scenes to make that possible. And like Tiffany said, one part of that is providing this scalable, cost effective infrastructure, which for us involves not just providing gpus and GPU compatible chipsets, but also innovating with our own custom silicon like tranium and inferential, because we will and truly know there can’t be any innovation if the cost model doesn’t make sense, and also performance and low latency. So that’s one aspect, and I think we mentioned a few times today as well, skills. Imagine if suddenly businesses have to have highly skilled data scientists who are phds and experts in AI. Surely there is a place for those skills. But not everyone should need to have those skills to be able to innovate with this technology.

Rada Stanic:
So our second sort of area of focus is on building tools and providing businesses with tools that they can really use to build these generative AI applications without having this deep skill set of these types of individuals. Another thing which we always say is security. There’s zero job for us. So how do you now, with all of this power that generative AI offers, how do you make sure that any business can really be fully assured that their data is not going to be used to improve these public models? Or also equally, how do we protect these model providers and ensure that the intellectual property is safe as well? So that’s another area. And then finally, also at what we call top layer of the stack, is how do we provide ready to go AI services so that customers have minimal worry about technology in itself. I mean, Tiffany just talked about call center and use of generative AI within call center environment. Now also imagine building on top of that and saying, oh well, now I’m going to give you an AI expert as a service that can actually be an expert within specific business. So you sort of point it to your internal data repositories and it can behave like a proper expert within the business.

Rada Stanic:
So you can see how the conversation goes from chipsets that we also make, but all the way to these high level AI services that provide an out of the box capability for businesses to leverage. So we try to, again, I think you’ve heard us say a few times, removing undifferentiated heavy lifting, hiding the complexities so that customers can focus on realizing that business benefit. That’s where we always start from, what are you trying to achieve? Is it a new and innovative capability? Is it cost saving? Is it improving something and then work backwards from that? What do we need to put in place to achieve that outcome? Which is why we can have customers like those who are creating their own large models. Like I mentioned, Leonardo AI or LG AI research has created their own foundation model, but not everyone will do that, especially maybe in, you know, lots of businesses who just want to say, I want to leverage scale and security of cloud to just create my own innovative applications. And I think we discussed one and Z Zero is a fabulous example of doing just that. And there are some in Australia as well. In a girl I always like bringing example of adore beauty, summarizing customers reviews of products and providing greater insights, how one product is kind of stacking up against the other. So it doesn’t have to be something mission critical, it can be something very simple, but it adds value to the business and the end user.

Paul Spain:
Now, when we look at the New Zealand market, we’ve got large and small organizations and every variation in between. Some of these are very small. Just an individual person is a business, right? And often having conversations with some of those that are in smaller businesses trying to work out, you know, where do they, you know, where should they leverage AI within, you know, within their business. And of course, there’s sort of off the shelf things like chat, GPT, copilot, and I’m sure AWS has things that sort of fit into that space as well. So there’s that sort of tool. Then there’s specific platforms they might be using, like Xero, which has over time been building in varying levels of AI. Be it to scan receipts and save you having to manually type bits and pieces in there, or ingest invoices to some of the newer staff where you’ll be able to kind of have a conversation with zero and fire your accountant. No, the accountants don’t want to hear that.

Paul Spain:
But that ability to be able to have these conversations with a platform or for that platform just to do heavy lifting for you, that maybe hasn’t been able to do in the past. I think Xero is one of your customers. Canva is another one of these kind of tools that’s often used within the small business and they’ve moved down that AI track. Is that kind of the direction that small businesses should generally be taking as looking at those off the shelf type platforms that are incorporating AI? Or should they have a bit more ambition around looking and doing more with AI? Any thoughts on that?

Tiffany Bloomquist:
Yes, I do. Don’t underestimate some of these small businesses, my friend. I was actually at Canterbury Tech doing a keynote down there, and one of the examples I gave was of a company that is actually leveraging a small company, brand new startup, leveraging facial recognition for sheep, because it’s very important for farmers to actually understand which of the use has how many lambs, and then to use that data to actually think through their breeding cycles. Okay, so the application of technology, and that’s literally using facial recognition with machine learning behind it, like all that stuff is so pervasive that I think we’re thinking about AI as a separate thing on its own right now. That’s its own thing. It’s just going to be in the products that we use. It’s just going to be in the experiences that we have. So what I think I would deeply recommend any small business do is know what is specific about the service that you provide.

Tiffany Bloomquist:
That is your unique IP, your unique special factor, whatever that is. That is core to what you do. Figure out how you leverage software and technology to amplify it, but stay true to that core and keep learning. One of the things that I definitely have to land in this conversation is the amount of information and training available on AI is massive right now. I mean, AWS just has a foundation in putting a lot of skills out there. So there’s like 600 free training courses you can take with AWS educate today, or you can use our AWS skill builder if you want to do little tiny bits. But we launched a whole AI ready program, which again, is free content for people to learn about what types of scenarios AI is meant for. I just think it’s going to be ubiquitous.

Tiffany Bloomquist:
And so the more people understand and aren’t afraid of the concept or seeing as this weird, scary big thing, the easier it will be. At the end of the day, though, it is about data. It is about organizing your data and understanding your data well enough to be able to put it together and create an outcome for a customer, for an employee. That is one of the number one structure pieces that organizations are working on right now. So even before some of them can start doing AI, they’ve got to get organised in where their data sits, how it’s structured, how it’s governed, how their operating model actually supports that. So I think while people might feel they’re, quote, behind in the AI game, most people are investing in truly understanding that data, how to leverage it. So one, get the skills. There’s a ton of free stuff out there.

Tiffany Bloomquist:
We’ve got tons of it that’s available and people should be leveraging it. And then two, figure out what is unique about your business, that data that you’re going to hold, how you’re structuring that so that you can leverage it later on down the line. A great example of AI and ML. We’ve actually been using it for ten years as Amazon. If you think about the ability for us to ship products anywhere around the world in a matter of weeks, and in the US, it’s days. It’s literally like two days. The only way that we can actually understand demand for product and forecast it, the majority is done with AI. It is just seamless to the customer.

Tiffany Bloomquist:
But they get an amazing experience by getting that product when they need it, having it ready in a fulfillment center that’s near where they live, that will just improve over time as more and more companies see it as ubiquitous and embedded in, versus something that’s just an add on or something that’s separate.

Paul Spain:
Lots there to.

Tiffany Bloomquist:
I know, sorry, I won’t keep going.

Paul Spain:
We’re probably out of time. And I know there is definitely more that we could get into. Is there probably anything that you can point listeners to in terms of really any resources or tools that haven’t been mentioned already that you think this is really important?

Tiffany Bloomquist:
Maybe we start with anything you want to share.

Rada Stanic:
It covered it really all. I think one of the biggest things for all of the customers of all sizes and industries is some of their training and the platforms like AWS skill builder, where we provide sort of Persona based training. So whether you’re an executive or you’re a developer, or you’re a business analyst, you can pick the training that best suits your specific role. And I think that’s key. And the second thing I would say is, for anyone who is really serious about generative AI, people think that some of these large language models, that they are magic. And the reality is they can only be as good as the data underneath. So I think businesses focusing on perhaps building their modern data platforms, looking how they govern the assets that they have, how they prepare them, clean them, and have them ready for use by AI, is something that’s not that well understood, I would say, because that is going to guarantee the ultimate success of generative AI. And we do a lot of free training also in person, virtually.

Rada Stanic:
You will see us advertise every now and then, announcements for all sorts of immersion days and hands on workshops and enablements. So we really invest in educating community and customers at large so that they can focus on what is it that I’m trying to solve, and then we sort of help them solve it.

Tiffany Bloomquist:
Beautifully said. I’ll just add on. There’s also a bit of a kiwi flavor in here. So one of the things that we do is while there’s many courses online available, anyone can have access to that. We also realise that there are times when people really need to invest in the communities that they operate in. So we’ve got several programs. One is called Hippori Wahine, which is targeted to create a community of women. And it’s a self paced learning.

Tiffany Bloomquist:
To get your first AWS certification open to women, and it’s funded. So for them, it’s free. And they have an opportunity not just to do the learning, but to actually be involved with a whole set of women that are invested in their own skills and trying to figure out what else is out there for them. Especially those that need a self paced learning, because they’re either brand new moms or moms looking at going back to work or thinking about a different career, they’ve got another job. They’re trying to understand what that looks like. It also means, though, that we have a responsibility to make sure that skills are available to everyone. It’s a part of why I’m wearing my dream house shirt. For those of you that can see it, if you’re on the podcast, I’ve got a gorgeous black t shirt on, and it basically is promoting a program where we actually have an interest in helping to educate the community around what it means to be in tech.

Tiffany Bloomquist:
That it isn’t somebody in a hoodie in a corner in front of a laptop coding. It can actually be a wide variety of things. And that technology should be embedded in all of what we do and that it’s a fantastic career, regardless of what you look like. We often host these dream house events at a marai or at a local skills center so that the entire community, grandparents, kids, can all see and be encouraging folks on what that journey looks like. So for businesses, get educated, get involved. For people. Also know that a part of our report from access is when you’re leveraging digital skills, there’s actually a pay boost as part of that. Oftentimes we’re seeing, I think it was 30% that was listed of an increase in expectations around pay because of leveraging digital skills in your job.

Tiffany Bloomquist:
So we want everybody to know it’s a positive thing from a technology perspective, but do lean in, get certified, get trained, understand how it can impact your life.

Paul Spain:
Yeah, that’s a really interesting number, the 30%. And we probably don’t have time to delve into, have to read the report into that, but we’ll look out for the details in the report. So that’s really encouraging. And I think it is important that we keep encouraging family, colleagues and so on to double down on building their digital skills. And obviously, AI is increasingly an important part of that picture. Well, thank you both for joining me on the New Zealand tech podcast today. Of course, also thanks to our show partners One NZ, 2degrees, Spark, HP and Gorilla Technology.

Paul Spain:
And yeah, thank you, Tiffany. Thank you, Rada. It’s just been a real pleasure to chat with you today.

Tiffany Bloomquist:
Thanks for having us.

Rada Stanic:
Thank you so much.