\We're diving into data with Snowflake's New Zealand country manager Tony Shaw.
Shaw shares how any size of organisation can benefit from data analytics and AI, emphasising the importance of having high-quality, consolidated information to drive business outcomes.
He also talks generative AI and the role that good data practice has in realising its potential and mitigating risks.
Plus, tech under Trump: the influence of tech billionaires in the recent election, what it might mean for tech policies, and how it may impact NZ's tech industry.
The Business of Tech is sponsored by 2degrees for Business.
Reading list
It is now clear that we've achieved the most incredible political there.
Look what happened?
Is this crazy? But it's a political victory that.
Our country has never seen before, nothing like this.
Look what happened. Indeed, Donald Trump is heading back to the White House after a stronger than expected showing in the presidential election, with the Senate and the House of Representatives destined to also be in Republican control.
Which has major implications for Trump's policy agenda, including a host of tech created issues from AI regulation to cryptocurrencies, data and privacy law reform, and the tech arms race with China.
This week, on the Business of Tech, powered by Two Degrees Business, we look at what a Trump administration means for tech, rold at influential tech billionaires and platforms played in the election campaign. I'm Peter Griffin and.
I'm Ben Moore. Coming up on the show as our featured guest. Tony Shore, the New Zealand country manager for Snowflake, a company that isn't as well known as Microsoft, AWS or Google, but is working with those companies and a rapidly growing roster of Kiwi companies to help them store and manage their data, run data analytics and use AI.
Yeah, Tony has some great advice for companies eyeing up data, analytics, machine learning, and AI, which is really about the importance of getting your data house in order before you delve into these things. So stick around for Ben's interview with Tony. But first we need to sift through the ashes of Wednesday night's election results. Ben, we weren't actually going to talk about the election on this episode because normally we record the podcast on a Tuesday afternoon, which was just before the election. Terrible timing, but yes.
But I happen to you got quite sick on Tuesday night, so we pushed the publication of the Business of Tech out a day, which is why you're hearing this on a.
Friday, Yeah, which sort of means we can reflect on what went down and look at what Trump may have in store on the tech front. And you know, I think before we get into that, we should really talk about the influence that tech billionaires and their platforms have had on this presidential election. I mean it's pretty clear in his victory speech from Florida, Trump very much talking about Elon Musk. You know, last time around, Elon Musk was relatively close to Trump. I remember that iconic meeting where he basically called in all the heads of the tech companies to Trump Tower and sort of had a chat with him. Peter Thiel was there, I think Tim Cook from Apple was there. Elon Musk was there, so it all seems sort of quite Trump appointed him to a couple of advisory councils and Musk ended up quitting them, so there was a bit of a falling out between him and Trump. He's really rekindled that relationship, and you know Elon Musk, I think he sees as being quite key to the success that he's had, so he'll need to repay Elon Musk. Just talk about Musk taking some sort of role the Department of Government Efficiency DOGE, so I think this will be pivotal. You've also got Musk clearly has his fingers in so many different pis, so many different companies, so lots of scope for conflicts of interest there. And of course with X and I think we saw this definitely the morphing of X in the last year or so during that campaign really into a conservative stronghold that's also been influential. So all of these things are intertwining. I think in Trump's favor.
Yeah, it definitely has been a swell, a more public swell of support for Donald Trump from that kind of tech elite in the last over the selection. And I think it speaks a little bit to the way maybe that Trump does approach these these kinds of companies where he wants their favor, He wants them to kind of be on his side, and as in a return, he'll be on their side. And we've seen him, We saw him use these kind of anti monopoly laws to kind of put pressure on the ones that maybe weren't as vocal of support for him in the past. Not to say that there's a direct link there, but it's hard to not see some kind of correlation at least. So Elon Musk, with his ambitions for Tesla and for SpaceX and for a lot of the defense contracts, must be really glad to be in Trump's good graces now, especially because Trump has talked quite a lot about increasing the amount of private companies contracting to defense and working and spending a lot more money on defense, which will mean more money in the pockets of these tech companies working on defense technologies.
Yeah, it's got to be good for SpaceX, I mean Tesla. Trump has been very anti electric vehicle, so will he now pivot to suddenly being influenced by Elon Musk along the lines of, well, no, EV's actually makes sense. I'm keen to support them. Whether there'll be more subsidies for evs, and likely given Trump's interest in the liquid gold, you know, the energy industry, the fossil fuel industry, but definitely on SpaceX, very much in bed with the US defense, so there'll be synergies there. And as yea, his other business obviously x AI and Grock building, you know, the biggest AI centric supercomputer stuff like that. I can see Musk sort of putting proposals to Trump about what the government should be doing with AI, and I think on a lot of these issues, Trump really he's not a tech guy. He doesn't really get this sort of stuff, but he's very influenced by the people around him. So you've got Musk there as a trusted advisor. You've got jd Vance, who has a previous life in venture capital, worked with Silicon Valley to fund companies, so he's very much embedded with some of that tech elite. You've got Peter Thiel who's a big backer of Trump as well. New Zealand citizen, founder of Pallenteer and AI company that does a lot of work for US agencies, police force, and military. There's share prices surging at the moment on the back of their latest results, all driven by AI. We've got this sort of cluster of people in Trump's orbit who have strong ideas about where the tech world should go, and he's listening to them and he owes them. A lot of them put money into these super packs to get him into power. They didn't raise as much money as Kamala Harris did in the Democrats, so it was all a bit of a waste of time and money. But he now has a lot of bills falling due, and what is that going to mean for the flavor of his policies. That's going to be the big question.
Yeah, I think a lot of it is going to central around deregulation. Personally, I think that's going to be a big flag that Trump will be waiving is getting out of the way of these particularly those Silicon Valley companies. I think the likes of Google and Microsoft may see some continuation of those anti trust kind of approaches. But when it comes to the those on the cutting edge with AI and the ones that are kind of more and Donald Trump's in a circle, will start to see that a lot of deregulation there seems to be pretty explicit in what he's been saying. And the same goes for cryptocurrency as well.
Well. It's interesting on crypto. You know, Trump has done a bit of a U turn. He was quite sort of hawkish against crypto a few years ago, and again I think this is people getting to him, and probably Musk, who you know is a big fan of dogecoin and is a crypto advocate as well, basically saying to him, no, you need to support this. He's done a U turn. He wants minimal crypto regulation. He'll probably limit the rather hawkish moves by the SEC in the US to regulate cryptocurrencies and those digital asset markets. So that's all well and good, but is that going to lead to more of the sort of FTX style implosions that we've seen. He won't want that either. On deregulation, sure, last time he cut tax and red tape. Businesses love that. It means they can spend more money on R and D and return more money to shareholders. So any company, particularly those big tech companies that make a lot of profit, they'll love that. But it was Trump after all, that kicked off a lot of that antitrust stuff a few years ago, So yeah, will he continue that and see the breakup of Google and others. But I agree with you. I think you know, he clearly doesn't like a monopoly. He's a free market guy, that's his philosophy on this. But he does want to see all of the red tape and the restrictions removed from really innovative companies, and at the moment they're the AI one. So I don't see him carrying forward some of those executive orders around AI that Biden put in place. I think we he'll dial that back sognificantly.
Yeah. I think the other area where we're going to see a big retraction in the US at least is green tech. So if there were you know, we've had a lot of eggs in the green tech basket here in New Zealand with our startups, and that may indicate that the US is no longer a viable entry point for these companies to really scale. So maybe refocusing more on the EU. If the US was kind of a big part of your strategy.
Yeah, and you know there have been as part of the Big Reconstruction Act that Biden passed after COVID, there was green tech funding in there. So whether that will continue. One area that will continue which Trump and the Democrats are on the same page on as the semiconductor industry, the Chips Act. So Trump is very much of the view that we need more local production in the US of semiconductors, the really high end important stuff that runs ai to reduce reliance in the global supply chain on Taiwan, which is very vulnerable to attack from China. So he'll carry on that sort of stuff, things like five G. You know, he's expressed his dismay that a lot of that technology is provided by European companies. So again Biden was on the same page. And I think for Trump what it really all is about is taking on China and that continuing sort of pressure, whether it's through the form of taris or big tariffs on stuff coming from China into the US sixty percent tariffs potentially, which is quite staggering, but really that polarization of technology between the Western world and the Chinese world, and We've seen China in the intervening few years since Trump was out of office, building its own operating systems, trying to generate higher capacity semiconductors to go into phones and AI devices. Trump will basically accelerate that further by putting more export controls on the exports of high technology to China, tariffs and local productions. So I think we'll just see an acceleration of that.
And that's also going to roll over to New Zealand a little bit in terms of trade agreements. He's talked about getting rid of the Indo Pacific Partnership trade Agreement, which would impact New Zealand. So how that will interact with New Zealand's tech exports to the US would not one hundred percent clear at the moment, but you know there is potentially some impact there.
Yeah. And the other sort of local angle I guess is orcus. You know, this agreement this packed between Australia, the US and the UK really about submarines, but you've got orcust Pillar two, which is about other advanced technologies like AI, like quantum computing, stealth technologies, you know, high end military stuff. And I've been quite supportive of the idea of New Zealand being involved in Pillar two. Not necessarily around nuclear submarines or anything like that, but Pillar two, these advanced technologies, we should have a hand with our allies and developing those. And I think, you know, Orcust has run into some trouble. I mean, this submarine deal is so vastly expensive. Whether it will actually come to fruition is anyone's guess. I think the Australians are starting to realize what they've signed up for is massive. But the other you know, you've got South Korean others Japan are saying, hey, we want it on Pillar two because they're starting to see some of the stuff that the Americans, the Brits and the Aussies are working on and saying, you know, we want to see that the table in developing that stuff because there is a greater threat from China, so let's work on this together. Trump will just carry on thinking, I think around orcus he sees that as a way to shore up support military support among allies in the Pacific. Whether that will encourage New Zealand to join or maybe will there be more pressure with a new you right leaning ambassador in this country, Will there be more pressure for New Zealand to actually put its cards on the table and join Orcus. I think that's a possibility.
It's hard to see exactly where we're going in terms of the ramifications. You know, with the potential for a Harris government, it was a lot more of the same, but a Trump government because his rhetoric can be quite inconsistent. You know, there is some stuff that we can guess about, but at the end of the day, it's really going to be just reacting as things happen. So it's going to be really important to actually pay attention, I think, to what is actually happening rather than what is being said through Trump presidency, And if.
His last stint as president has anything to go by, it'll be those key personalities around him because he really is a bit of an empty vessel in terms of his thinking on some of these issues, particularly around technology. Now, just listening to him explain on election night, you know, the starship returning to Earth, you know, when he was praising you on Muscus, just clear he doesn't really understand this stuff at all, which is fine, but it's the people around him and the worry I think in the US at the moment is you know, this paranoia about the deep state in the US, this shadowy sort of left leaning cabal that runs America that he's been trying to root out. Is he just going to replace people at the SEC, his top tech advisors, people responsible for climate change policy. Is he just going to replace them with political appointees who don't really care about the evidence or the science or what technical advisors suggest is the right thing to do. And he's just going to take the advice off a small group of very wealthy, right leaning tech elites who he owes big time because they helped get him into office, you know, judging by past performance, that's what he tends to do. He surrounds himself with people who are loyal, but people he also relied on to get into office first time round. I think we'll see a lot more of that unfortunately.
Yeah, it really is about an exchange of wealth and favors and keys and power and deals. It really is all about the deals.
It's transactional with Trump, and people have advised you know, if he does get in, you've got to treat it as a transaction with whether you're negotiating what to do with Ukraine or a trade deal with a country. It's transactional. You need to be in that mindset dealing with this guy. So maybe that's the approach that maybe we should take as well. Absolutely, so clearly it's going to be an interesting year. Head We'll keep you posted and give our analysis on everything related to tech as the Trump administration settles in. But Ben, whenever we interview companies around some of these issues like AI, predictive analytics, and all the cool things businesses can technically do these days with the data generated by the businesses, we get the same surprising response.
Yeah, the conversation typically grinds to a halt, and we are told that a lot of our businesses just don't have their data in the right places, in the right formats to do any of that.
So talking about it's a bit of a waste of time if the basics really aren't done well.
Which is why we're hearing a lot more from companies like Snowflake and data Bricks, companies that have emerged in recent years to help organizations manage that data.
They're basically data warehousing and analytics platforms that try to get all your data in one place, process it in a uniform and secure way and interact with the various applications you're using to run your business. I was actually staying at a hotel in Awkant recently and found myself walking into the middle of a Snowflake conference. It was actually quite a big affair.
Well, data is a big affair now, it's big business and Snowflake has around two hundred customers in New Zealand to date. It did around one hundred million dollars in revenue last year just across Australia and New Zealand, according to its financial accounts filed with the company's office.
And spending on data, warehousing and platforms is really growing fast. So Ben, this is a timely interview with Tony Shaw, who's been around a tech industry for a long time since at NCR, dell, IBM, MuleSoft as well. Let's listen to your interview with Tony Shaw and come back for some thoughts on the back end.
Thank you so much, Tony for joining us on the Business of Tech podcast. It's really great to have you here. Why don't we start with just a little bit of background about who you are and what you do.
Oh fantastic, Heyn, Thank you so much for having us on board today. My name is Tony Shaw. I'm the country manager for Snowflake in New Zealand. I've been with the company just coming up to six years now, so quite a long time to be with one organization. But I've always been in tech and for a long time in analytics. I originally started my career working for NCR as a financial analyst and pricing and planning and using data and realizing how important it can be to make financial decisions. And then from there I moved into more of the sales and business development side of things, both in New Zealand and I had a long time in London, and then predominantly in the data and analytics side of things.
Do you want to share maybe the perception of data maybe pre your Snowflake time, and then how it's changed since then?
Yeah, no problem. I think data's always been important. Organizations have always had the aspiration to be using data better to make better and informed decisions. But what's happened in the last five to ten years is the accessibility of the information has become so much easier. The cost to get that data and analyze it has dropped significantly, and that's opened up massive opportunities because it allows organizations to bring all of their data from all of their disparate systems into one environment where that structured data unstructured data, and then can imply analytics to that. Historically, it used to be a lot of backwards looking, a lot of reporting what did happen? And now where we're seeing is a lot more predictive analytics, opening up the information to a lot more of the business users and allowing that decision making to be a lot more in the front line rather than just the back office. So we're seeing that dissemination of information across multiple channels, multiple users, and the ease of use. So it's no longer just a back office function as lines of business making decisions every day which are moving the dial within those organizations.
Right. And you know, traditionally when we think of data, we think big data, right especially these days, and we think big companies. But that's increasingly changing as well. I would imagine, like you say, as the accessibility, the affordability of data and data analytics tools are starting to shift a little bit, are you starting to see smaller companies, you know, not necessarily your one person companies, but maybe your medium size businesses gaining a better understanding of how to utilize their data.
Yeah, it's been remarkable. Since we started the business in New Zealand in twenty nineteen, we were looking at what is the segments and what is the segmentation and customers that we're going to look to try and require. We had two stomers in New Zealand when we started the business here and now we've got north of two hundred, and it was really interesting. We started to think around that segmentation and we thought it might be some mid tier customers and then you know, maybe we can work our way up or down across the different spectrum of size and scale and complexity. But what happened was we've got organizations of all size and scale very early. And I think one of the things with Snowflake, and one of the reasons why we had such fantastic adoption, was it's the ability to scale down to New Zealand size companies, not just being able to scale up. So there's the global organizations, you know, there's the capital ones and the sinespres etc. But within New Zealand because the platform scales down and you only pay for what you use on a true consumption basis. We've been able to scale to organizations that are getting enterprise enterprise grade capability, but they're only paying for what they use based on the size of the organization or how much they actually need to use of the platform.
To what extent a New Zealand companies really using all of the capabilities of Snowflake. Are we up there in the most advanced users or are we kind of just using the very basics because we're smaller.
And yeah, absolutely. We just had a conference last week. It was unbelievable. We had over a thousand people there, which makes it the largest data and analytics event in New Zealand, and we had some fabulous customers. So we'd organizations like in New Zealand, Tavado, Aura in zed Health, in zet, super one, end, z MITA ten, Spark, shares e'se the kind of list goes on and it was a really great opportunity for organizations to share what they're doing and how they deliver value from the platform, and also to build that community so organizations can network with their peers and learn from each other. But in terms of taking on the global stage, shares Y's is one of our fantastic customers. They actually recently won the APJA Data Driver Award for powered by So what that means is they're powering their application using Snowflake to help drive the adoption and understand their customer behaviors in order to deliver a better service. And they've just had phenomenal growth. So you know, they've got seven hundred thousand customers. So we're taking on the world. Shares is being successful across here and across in Australia, and we've got a number of tech startups that we're working with who are winning awards and delivering really fantastic results for their business on a global scale.
Fantastic. Yeah, so it sounds like you've got some real power users. Then that's what you're.
Saying, unbelievable. It's both the business users. So we had the co CEOs presenting around how that's driving value. Data analysts we have technical capabilities. It's the ability to work with all of the different personas across an organization, not just the technical people though they love the platform, so the architects, the engineers, the really deep data people, but then also the people who are consuming it, so technically literate business analysts people just writing natural language questions in English, executive writing. Sorry, just analyzing what's happened and what's going to happen and their business. That's across the board, those different personas that all use data in a slightly different nuance, but they want consistency of information, they want high quality, they want real time, they want accurate information so they can make those decisions.
Cool. Now, obviously you can. It's great to talk up to customers that are doing really awesome stuff, but New Zealand's definitely not perfect nowhere is in terms of how it's utilizing data. So what are some of the areas that you're seeing New Zealand lagging behind? Maybe some New Zealand companies where you think, you know some areas of focus could be to improve the usage of data within New Zealand.
I think that the pitfalls that we always see is making sure that there's executive sponsorship and outcomes that the organization is trying to drive towards. So it's really important that it doesn't become a science experiment or a program that's just for the IT users. What the successful organizations do is they've got very strong alignment to a specific business outcome, whether that's a finance program looking at receivables or finance transformation, whether it's marketing looking at customer experience, NPS, churn, cross sale, etc. Or operations to streamline the efficiency with which the organization works in it has to have that business outcome that everybody can anchor themselves and align to. When you've got that, that goes a long way to making sure this program's success. And then the usual governance across the program and making sure that there is steps along the way that people are measuring to make sure that that outcome happens. When you start to get those sorts of things, then everything else just falls into place.
Cool. Now, let's say I'm one of the New Zealand companies that hasn't started to get deep into data yam. You know, maybe a medium sized company who is starting to think about the potential there. What are my first kind of steps?
The first step is defining what dial within the business you're trying to move and what is that outcome you're trying to achieve. So say it's a marketing outcome around cross seal. Make sure that those objectives and those metrics are well understood and documented, and then start small and try and deliver that program so that you deliver that specific outcome, get the win, and then build upon that. You need to paint the vision to the organization in terms of what is the analytic capability going to deliver. So you need to have a vision and where we're going as an organization, but you also need to have a specific outcome that you're driving towards that you can build on that success. Then you need to drive where do I get the data from and how do I get high quality information to solve that business problem and answer the questions that you're looking to define or answer sorry, And then it's getting the technical teams aligned to find that data, source that data cleanse that's high quality decision making because you want to make sure sure that the information that's being used is of quality so that the decisions out the back of it are influenced.
What does that mean cleansing data? Like, what does that actually in real terms mean? Because if I'm a company that's been around for twenty years, I've got a bunch of spreadsheets and PDFs and all this kind of stuff and it's ordered, it's in folders. We know where everything is. But is that clean? Is that clean enough.
It depends a lot of the times those spreadsheets have been built up by a couple of specific people. They might be suitable for that use case or that specific piece of information you're looking to deliver. An example, when I was a pricing analyst, we used to have huge amounts of spreadsheets everywhere that have interconnected links, and we put out some pricing models. Then you'd come back about a month later and change something because you'd found a mistake in the spreadsheet in the formulas, and that would change the entire pricing model. And you'd be sitting there going, oh my goodness, now I've just completely stuffed this up. You change something else to get it back, and then the numbers would all change back again. Spreadsheets, whilst they're across every single organization, are kind of the bane of any enterprise organization's life because there is no real auditability. So when I talk about high quality data, it's getting that data from those source systems, making sure that it's usable and in a format that's understandable, but it's consolidated across multiple touch points so that you've got a consistent view of customer, and then involving the business teams to define what is the rules and logic so that everybody knows what the definition of a customer is, what is a definition of revenue or profit or what happens to be, and then everyone's working off that consistent set of information. We've all been in meetings where people are arguing about the data rather than what they do with that information. So what we want to try and do is consolidate the information so that it's a single view across the business. And then people are thinking about what are the decisions I make, not hey, is that the right one? Am I questioning the actual data validity rather than what I can do with it?
What's the kind what's the kind of talent that you would need to do that? Do you need to hire an house data scientist? Is it okay to just kind of get a consultant into kind of do some data stuff for you to get you ready.
I think consultants have a place and they've got a lot of experience that can bring to bear on organizations. But I think the organizations themselves have a responsibility and they have to have a capability internally. This can't be done to an organization. You have to do it with the organization and the people within the enterprise or the company. They know what the business is trying to achieve, they know where to get the data from. So you need to have a set of skills within the organization, and that skills from a technical capability to work out where does the data come from and how do I get it and then also how do I put that into the hands of the users so they've got confidence that they can start to drive analysis from it. But the internal capability is critical. One of the things we're trying to do at Snowflake is build a really big community. So the event we just ran with a huge number of people. We run user groups, we run meetups, we run product specialist workshops. When we bring some of our teams offshore into New Zealand, and it's really important to build that community and network so we can share what's working and to your point before, what's not working, so that we can avoid those pitfalls where possible and start to accelerate how do we deliver that outcome. But I think internally the capability needs to be there. You know, we need to train our teams, we need to cross pollinate from existing teams, so you might have somebody who's working internally within an organization, they've got a huge amount of tribal knowledge within that organization. But then how do we cross pollinate their skill sets with whatever it is they need with it's technical or analysts. So they've got the data literacy to drive that outcome. But absolutely internal was critical.
Yeah, I mean, I guess the message that I'm getting really is that you can't do this. Lais a fair. You can't just be like, let's dabble in some data. You really have to sit down and create a cohesive, strong plan and roadmap and objectives and spend the time to actually build that out. And if you're seeing gaps in your organization, then you actually need to maybe fill those gaps, whether that's with training or with bringing on new staff. Does that kind of sound about right?
Yeah, one hundred percent agree. And it's building that strategy into the business outcome of the business strategy, so that the data strategy is part of your business strategy because they shouldn't be separate. One can inform the other and the other can form each other. A lot of times the data teams have a really enterprise view of the business because they're looking at data from multiple different areas, so you're not siloed within say HR, or siloed within finance, or siloed within marketing. The data teams get a strong visibility across the organization. For example, we were talking in New Zealand presented recently at our conference. Again they talked about the concept of majors and miners, So you've got a data team which has got majors in data and analytics, but working with the lines of business who have a minor and data, but a major is in their skill set, whether it's HR, where it's cargo, whether it's financed, whatever happens to be. And that dovetail together of the data literacy and the data capability with the knowledge of that specific line of business and what's important to that line is really important because then you're marrying both of the outcome and the capability together, which drives a lot of value for the organization. Having that strategy which is aligned into the business strategy is really important. And obviously with the introduction of things like AI, AI is built upon how you use data, whether it's internal data, external data to drive that decision making. So you know, you can't really have an AI strategy without a data strategy, and all of this should be blended into what is the objectives of the organization and what are they driving for?
Great, you just did my job then, and you pivoted to exactly where I wanted to go next. We just talk about AI because we have to, right because it's twenty twenty four and the last three years of AI has been just a different story completely to where it was previously. From your experience as somebody who lives and breathes data, what has that experience been like to watch data go from this kind of dry but necessary thing to the thing that is powering the future.
It's absolutely amazing, isn't it. You know, you know, in your personal life you use things like GPT and the output that it can deliver is just extraordinary. And where it's going, I think is fantastic. It's outstanding. But you've got to get the foundations right because otherwise you're building on quicksand and you're analyzing inefficient data and you'll very quickly lose confidence from those users. And there's also some of those traditional techniques which are still incredibly valuable to the organization. So just making sure that we understand the vision, and we go after that, and we go after that with speed, but at the same time we don't take the focus off some of those other areas which we can deliver very very quick value to the business.
Somebody said to me the other day that actually, with new AI models, data cleanliness is actually not as important as it used to be, because if you look at something like chat GPT, you know, the training is not necessarily there's so much of it they couldn't possibly go through and clean it all. Do you think that's true that if you were wanting to create kind of a GENAI model that's used you can use to analyze certain levels, certain kinds of data within the organization, that there is less need for data cleanliness than they used to be.
I'd say, where you're looking at the entire Internet for consumers like you just describe, maybe when organizations are looking to use information which is internal to their organization, that has to be very high quality. It has to be robust, it has to be trusted, and it has to be using the information and knowledge from that organization to prevent things like hallucinations and bad decisions because the data is incorrect, inaccurate, not full enough. There's not a quorum of data to make an informed decision. I think the data quality aspects are even more important for an organization using some of these advanced capabilities like genai. The Genai capability helps a lot in terms of being able to put some of that tagging, say, or definitions around what some of that data means. It helps speed up the efficiency to make the data more reliable and higher quality and understood. But without putting the thought into having that high quality data, it's going to fall flat. In my opinion, I think we need to absolutely focus on the availability, the security and governance, the privacy, the quality and trusted data and then apply these techniques on top of it. And one of the things where Big believers on is bring the processing and the workload to the data rather than pushing all the data out to different systems. And the reason for that is because you've got that single view of the business, you've got one place to make sure that the data is of that high quality we're just describing, and you've got the privacy and governance so that only the right people are allowed to see it. Because what we're doing is we're opening up the access to a huge wide range of different consumers of the data, So we've got to make sure that it's protect and we've got to make sure that it's of high quality.
Yeah, you're not on the difference between creating something from a mass market and creating something that is to improve organizational performance, and those are two very different goals completely. The other thing that I've been considering about generative AI lately is there's this kind of to and fro about how much we let the GENAI actually do if that kind of makes sense, where it can be quite creative and thoughtful and very have high contextual understanding, but that may potentially, you know, open up the hallucinations or we're not quite sure where that's going to go. Or we can be very tight and strict and be like, it can only return these information from these sources in these modes, and trying to find the balance of that can be tricky at a kind of data level when you're figuring out what to include and what not to include, how do you start making some of those decisions.
You're absolutely right, it is a tricky decision or tricky consideration to think around. Where we think around it is having access to the right amount of information, but putting those governance and controls on there so that you've got things like role based access so that only I'm allowed to see the information that's purten in to my specific role and I can't see anything else outside of that. And that's why it's really important to get that governance and that privacy foundations set and defined upfront so that it's not being made up and make sure that you've got the right level of data to support the decision that you're trying to solve. And my view would be start small, start to prove out some value, and then expand as you've got that confidence within the business. But it's moving incredibly fast, right you know. You think even a couple of years ago, you know, Chat GPT was just coming of age and people had only just staid to hear about it, and now AI is embedded into just about every single platform and process. What we're looking to do is understand how we can use each of those different silos of informations and applications bring that together so you've still got that holistic view at the data level, not just at the application level. So you want to be able to bring that data together from multiple places and then apply AI across it, depending on what it is you're trying to do, but you know it's moving so quickly it's really exciting. To be parely honest.
What is exciting about it for you? Because you know, for office workers there's that kind of productivity gain stuff that's being talked about. For consumers there's like access to information that they may not have or ability to proof read and do these kinds of everyday tasks. But as somebody who is like super deep in the world of data, what is actually super exciting for you about the generative AI stuff?
The productivity part that'll be part of it, But I don't think that organizations are looking at just the productivity. Sure there's efficiency, but I think it's the upside that people can drive from it. Is the better network planning in TALCOS is the better supply chain management. Because you're pulling information from third party suppliers as well as the internal information. You can run and advance large language model across that to work out what is the route processing or where do you deliver things quicker? That outcome that's going to move the dial with those organizations to drive revenue or make them more profitable. That's really exciting and obviously the productivity gains will come as well.
Are we already seeing some of those gains in certain areas using the new air models? Like can do you have examples of that? Yeah?
Absolutely? Mine to ten was just talking. They've spoke to our conference again last year. They've had a very small team, so they've managed to consult it a lot of their information one year on what done as they've applied some of these large language models to look at water supply chain and how can they deliver better outcomes across the retail organization. So they're starting to embed some of these capabilities into their processes. We're seeing the talcos doing the same things. A lot of them have had proof of concepts that they're now starting to put into production. So I think that there's going to be a lot of the pilot and prototype pieces of we're really accelerating now, and to be honest, some of the organizations they see that as a competitive differentiator, so they are actually keeping some of them relatively close to their chests because the faster they can move, they're looking to leap frog the competitors.
If twenty twenty one twenty two was kind of the emergence and the testing and the seeing what could go wrong? You know, twenty twenty three and twenty twenty four has been about getting those prototypes and starting to see what can happen. Is twenty twenty five to twenty six is that going to be the acceleration time? Where are we at in terms of starting to really see mass adoption of this tech at an enterprise level, at a fundamentally restructuring level.
Yeah, I think over the next twelve to eighty months you will see a massive acceleration of that. I think those organizations that have done those that foundational work are in a much better position to be able to accelerate faster. So those organizations that have got trusted, high quality, consolidated information, then they're looking at what do they do to exploit it. They've got that quorum of data, and now how do we use it and exploit it quickly? There's still organizations which have yet to do that foundational work, and that foundational work is critical before you can start to exploit it in a really meaningful way. So I think there's going to be those that are ahead of the curve and have been ahead of the curve for the last few years are going to be able to accelerate quicker than those who haven't done that homework and done the foundational stuff. And it's no different to whether it's GENAI or large language models. Those organizations that have got that high quality data, they've spent the time to ensure that the lines of businesses have data literacy and data skills and know what they can do with the information to change the processes will be in a better position. So adding on top of that things like genai, it will allow those organizations to go faster. But it's just accelerated how quickly organizations can start to exploit it. I don't think it changes the fundamental that you've got to get the basics right and do that well before you can accelerate.
What would you say are the biggest risks that we need to be thinking about as we enter this accelerative phase.
I think the privacy and just because we've got the data, does that give us the right to use that data mentality? And I think we've got to be really considerate that most of these organizations is not their data, it's their customers data, So we need to really consider what is it that we're going to do with that data and make sure that it's doing the right things for their customers as well as the internal organization. So we've got to think around the privacy, the use of it, the AI governance and governance of the customer's use of it, and the permissions and things like that. So I think the accessibility is great, but just because we've got it doesn't necessarily mean we should use it in a certain way. There's going to be a lot of focus around the obviously security, privacy, and then also how do we just continue to evolve on that as well?
What do you mean by that.
In terms of as the technology moves so much faster, how do we keep up with that? And how do we think about the new use cases? How do we think around what is that business driver again taking it away from just a technology problem, what is the business trying to achieve ross that line of business, finance, marketing, et cetera. And how do we align to that outcome right? Otherwise we can spend a huge amount of money with science experiments that don't actually do much for the business. Another thing that will be important will be looking at the cost considerations, making sure that the whatever we're doing is aligned to the outcome so that it's cost and value tightly coupled. Otherwise, you know, they're not cheap things to run, so we need to make sure we've got the guardrails across it. So cost management is going to be efficient, going to be important, that the governance and privacy is going to be important, and that all ties back to what is what are we're using it for and what is the business value we're trying to drive out the back of it.
So get excited, but not too excited, and be thoughtful. That's kind of there, I think.
Be excited, but be thoughtful. Don't don't limit what you think you can do because you probably can. And it's exciting time to go and test some of these hypotheses and see how it works. So be excited, to be really excited, it's going to be fantastic next couple of years. But just be thoughtful about how you're using it and thoughtful about your customers.
So, if ever there was a man who lives and breathed data, I think it's Tony Shaw. He has clearly been in the industry for a long time, and his advice I think, while some of it isn't necessarily novel. It's the stuff we've been hearing for a while about getting data and order. I think that the way that he has put it really was very clear and concise and actionable as well, which is what I appreciated about the chat.
Yeah, he really talked about this transition into data and analytics, the importance of data in financial decision making, and for years, you know, we've been talking to New Zealand businesses about writing about it, and they're all up for it, and some of them are really doing that, doing really smart things with data, but we were a bit slower to the move to the cloud and getting data in order as part of that digital transformation. So a lot of businesses talk about this stuff, but are they actually using it? And when I took to them sort of off the record, they say, well, actually, know where we've done pilots, we're doing limited use cases related to data analytics and the like, but we don't have the data in the right shape. We need to build a data warehouse or a data lake. We need to standardize our data and that literally for some of them is taking years. So we've seen that's why we've seen the rise of snowflake and data Bricks and others. The big tech platforms can only do so much. It's really up to you, and there's a layer between the customer and the big platform. We're all of your data potentially is going to be and these companies are playing a really valuable role there.
Yeah, and the couple that we mentioned Data Bricks and Snowflake, and these are the ones that have really come out swinging and have shown that the value over and over again. And Snowflake listed on the NASDAK and has shown really great growth since doing that. So you know, its success is I think a good indicator of the value that it is offering to organizations globally. And you know Tony talking about the fact that it can scale up to these massive, massive international corporates, but it can also scale down to fit the needs of organizations and countries like New Zealand. And if we want to be the country that is using AI, that is using our data to improve our productivity, to enter the brave new digital world and kind of stay relevant on a global scale, then these kinds of products, these kinds of projects of what needs to be done on a bigger scale. And what Tony was saying about not doing science experiments anymore, right the time for kind of these doing science experiments over and over again. The small scale dabbling is kind of if you're still in that phase, you might need to put a bit of welly behind it and get on with it.
Yeah, yeah, yeah, I mean, I think his advice is sort of what we've heard, which is start small, don't necessarily go big bang, because if you've designed it wrong, suddenly it becomes a very expensive failure. So target a part of the business where having great insights into your data is going to really help the business. Start that experiment a little bit, then grow a bigger But he's clearly predicting a major acceleration of AI adoption over the next twelve to eighteen months as organizations do that foundational work. Trying to get ahead of the curve is a competitive advantage. Hopefully that message is getting through in New Zealand. We've seen so much research over the last year or so to suggest that we're a little bit behind the curve. But if companies like Snowflake can help accelerate that, because as you say, it scales down to medium sized businesses quite well, that's basically where New Zealand plays and a lot of those companies have been playing around with co pilots and chatbots and AI related applications, so maybe some of them have done enough work to actually in twenty twenty five and beyond make really good use of AI. And again what we've heard from others is emphasizing the importance of aligning sort of AI and data initiatives with business outcomes and having internal sponsors, people in the executive off the business, people on the board who are real champions for this. There's no point doing something where the CEO and the executive team is sort of saying how much is this going to cost? If they're not convinced if the value of investing in these sorts of platforms to the business, you've got a problem. They've all got to be on board. So thanks very much to Tony Shaw from Snowflake for his thoughts on the data landscape and what's needed to spur AI adoption.
We'll be touching on that and next week's episode two, and we have a panel of AI experts joining us to look at the year in AI, big developments in the technology, regulation and government's use of AI, and what may be in store in twenty twenty.
Five Show notes for the Business of Tech are in the podcast section at Business Desk dot co dot nz, where you can stream this podcast in full every week. It's also available from iHeartRadio or your podcast platform of choice.
Get in touch with your feedback and we'd love to hear your suggestions for upcoming tests too. You can email me Ben at business Desk dot Co dot and z, and you.
Can find both of us on x and LinkedIn, where you can follow the Business of Tech page for all of our updates.
That's it for this week. We'll be back talk AI and way through the election debris next Thursday.
We'll catch you in
Mm hmm