1 00:00:04,960 --> 00:00:08,080 Speaker 1: Welcome to the Business of Tech powered by two Degrees Business. 2 00:00:08,080 --> 00:00:11,680 Speaker 1: I'm Peter Griffin, back with Season four off the Business 3 00:00:11,680 --> 00:00:14,520 Speaker 1: of Tech. Thanks for joining me once again for a 4 00:00:14,520 --> 00:00:18,360 Speaker 1: weekly dose of tech through a Kiwi lens, making sense 5 00:00:18,520 --> 00:00:21,919 Speaker 1: of the big tech related issues affecting business, the economy, 6 00:00:22,280 --> 00:00:25,600 Speaker 1: and society. And twenty twenty six is off to a 7 00:00:25,760 --> 00:00:28,680 Speaker 1: very interesting start from a tech point of view. The 8 00:00:28,680 --> 00:00:32,120 Speaker 1: big tech story over the holiday period was the hacking 9 00:00:32,720 --> 00:00:36,760 Speaker 1: and ransom attack on patient portal Manage my Health. I'm 10 00:00:36,840 --> 00:00:40,879 Speaker 1: a user of Manage my Health, but thankfully it appears 11 00:00:41,200 --> 00:00:44,839 Speaker 1: didn't have my data released onto the dark web. At 12 00:00:44,920 --> 00:00:47,159 Speaker 1: least I didn't get a notification to say that my 13 00:00:47,280 --> 00:00:50,879 Speaker 1: data had been compromised. It doesn't get much more serious 14 00:00:50,920 --> 00:00:55,400 Speaker 1: than sensitive health data being breached and stolen, So that 15 00:00:55,520 --> 00:00:58,920 Speaker 1: story has serious implications not only for trust in Manage 16 00:00:58,960 --> 00:01:03,360 Speaker 1: My Health, but our relationships with patient portals, particularly as 17 00:01:03,480 --> 00:01:07,959 Speaker 1: AI looms large in the health sector. Speaking of AI, 18 00:01:08,080 --> 00:01:11,000 Speaker 1: the big story so far of twenty twenty six is 19 00:01:11,160 --> 00:01:15,640 Speaker 1: confirmation that Apple will lean on Google's Gemini AI models 20 00:01:15,640 --> 00:01:20,840 Speaker 1: to inform its own foundational models. Siri and Apple Intelligence, 21 00:01:21,160 --> 00:01:25,119 Speaker 1: long rumored, now confirmed. A five billion dollar a year deal. 22 00:01:25,160 --> 00:01:29,039 Speaker 1: According to the Financial Times, it's a huge win for Google, 23 00:01:29,080 --> 00:01:31,959 Speaker 1: which was seen as losing the race with Open AI 24 00:01:32,360 --> 00:01:35,880 Speaker 1: on generative AI. That move seates it in the box 25 00:01:35,920 --> 00:01:39,080 Speaker 1: seat when it comes to powering the AI that billions 26 00:01:39,160 --> 00:01:40,560 Speaker 1: of people use on their phones. 27 00:01:41,200 --> 00:01:42,240 Speaker 2: Gemini, after all. 28 00:01:42,200 --> 00:01:47,160 Speaker 1: Is already the default AI assistant one hundreds of millions, 29 00:01:47,200 --> 00:01:51,800 Speaker 1: if not billions, of Android devices. So as usual in 30 00:01:51,840 --> 00:01:55,160 Speaker 1: the tech space, there's a lot happening. It's quite complex, 31 00:01:55,240 --> 00:01:58,000 Speaker 1: it's fast moving, and the premise off the business of 32 00:01:58,080 --> 00:02:02,040 Speaker 1: tech remains the same. In twenty twenty six, I want 33 00:02:02,080 --> 00:02:06,280 Speaker 1: to bring you interviews with interesting and informed people who 34 00:02:06,360 --> 00:02:08,800 Speaker 1: know a lot about technology to help you make sense 35 00:02:08,919 --> 00:02:11,040 Speaker 1: of what's going on in the world of tech, what 36 00:02:11,080 --> 00:02:14,720 Speaker 1: it means to you, particularly if you're in business and 37 00:02:14,760 --> 00:02:17,680 Speaker 1: these are mission critical decisions you have to make. But 38 00:02:17,919 --> 00:02:21,720 Speaker 1: for all of us who are using, consuming applying tech 39 00:02:21,919 --> 00:02:25,080 Speaker 1: in our day to day lives. This year, I'll double 40 00:02:25,080 --> 00:02:27,880 Speaker 1: down a little bit more on the idea of what 41 00:02:27,880 --> 00:02:31,639 Speaker 1: it means to us, unpacking what is driving the developments 42 00:02:31,720 --> 00:02:32,440 Speaker 1: in tech. 43 00:02:32,280 --> 00:02:35,119 Speaker 2: And crucially, how we should respond to them. 44 00:02:35,680 --> 00:02:38,959 Speaker 1: The case for a social media ban implemented as a 45 00:02:39,000 --> 00:02:42,640 Speaker 1: world first in Australia last month is, for instance, a 46 00:02:42,680 --> 00:02:46,440 Speaker 1: perfect example of an issue that needs careful unpacking to 47 00:02:46,680 --> 00:02:49,880 Speaker 1: understand the implications and what's involved. 48 00:02:50,600 --> 00:02:52,240 Speaker 2: I'm not going to chase the biggest. 49 00:02:51,919 --> 00:02:55,639 Speaker 1: Names in tech, the best known people, necessarily, I want 50 00:02:55,639 --> 00:02:58,560 Speaker 1: to talk to people who have insights and a way 51 00:02:58,680 --> 00:03:03,040 Speaker 1: of explaining tech that you experienced that AHA moment and 52 00:03:03,160 --> 00:03:06,720 Speaker 1: hopefully used what you've learned to inform your own decisions 53 00:03:06,760 --> 00:03:10,640 Speaker 1: about how you interact with tech to kick off in 54 00:03:10,639 --> 00:03:13,880 Speaker 1: that vein episode one of season four returns to the 55 00:03:13,919 --> 00:03:15,799 Speaker 1: topic of artificial intelligence. 56 00:03:15,840 --> 00:03:16,080 Speaker 2: Look. 57 00:03:16,120 --> 00:03:19,280 Speaker 1: I did about thirty episodes on AI last year. It 58 00:03:19,360 --> 00:03:22,880 Speaker 1: was the topic dajore and will remain so in twenty 59 00:03:22,919 --> 00:03:26,240 Speaker 1: twenty six. New Zealand loves to talk about being a 60 00:03:26,280 --> 00:03:29,320 Speaker 1: world class tech nation, but when it comes to AI, 61 00:03:29,520 --> 00:03:33,000 Speaker 1: the numbers tell a very different story. Just over half 62 00:03:33,040 --> 00:03:36,960 Speaker 1: of Australian and New Zealand organizations say they're using AI today, 63 00:03:37,200 --> 00:03:37,920 Speaker 1: compared with. 64 00:03:37,880 --> 00:03:39,880 Speaker 2: More than two thirds worldwide. 65 00:03:40,360 --> 00:03:45,120 Speaker 1: Only eight percent of our firms qualify as true frontier 66 00:03:45,280 --> 00:03:48,520 Speaker 1: companies that are using AI to transform how they work, 67 00:03:48,840 --> 00:03:53,400 Speaker 1: versus twenty two percent globally. That's according to IDC Research, 68 00:03:53,440 --> 00:03:57,520 Speaker 1: who surveyed four thousand business leaders late last year on 69 00:03:57,560 --> 00:04:00,920 Speaker 1: their approach to AI, and a study commissioned by Microsoft, 70 00:04:01,520 --> 00:04:05,240 Speaker 1: which with copilots and its Azure AI tools, has a 71 00:04:05,240 --> 00:04:10,040 Speaker 1: big stake in boosting AI uptake. That gap that IDC 72 00:04:10,280 --> 00:04:15,320 Speaker 1: identified really matters. IDC estimates AI could account for three 73 00:04:15,360 --> 00:04:18,920 Speaker 1: point seven percent of global GDP by twenty thirty, and 74 00:04:18,960 --> 00:04:22,640 Speaker 1: the frontier firms that are out there in front are 75 00:04:22,680 --> 00:04:25,839 Speaker 1: already seeing roughly three times the return on their AI 76 00:04:25,920 --> 00:04:30,159 Speaker 1: investments compared with slower adopters. If New Zealand keeps treating 77 00:04:30,240 --> 00:04:34,000 Speaker 1: AI as a fancy search engine or a personal productivity 78 00:04:34,080 --> 00:04:38,599 Speaker 1: hack while others rebuild whole processes and industries around it, 79 00:04:39,120 --> 00:04:43,680 Speaker 1: we risk baking our low productivity problem into the next decade. 80 00:04:43,839 --> 00:04:45,279 Speaker 2: So why are we lagging? 81 00:04:45,600 --> 00:04:49,960 Speaker 1: A mix of caution, skills, gaps, patchy governance and culture 82 00:04:50,120 --> 00:04:52,800 Speaker 1: is what it boils down to. We're curious enough to 83 00:04:52,920 --> 00:04:56,200 Speaker 1: play with AI on the side, but two risk averse 84 00:04:56,320 --> 00:05:00,440 Speaker 1: and underprepared on data, leadership and change to go big 85 00:05:00,880 --> 00:05:04,640 Speaker 1: in the core of the business. In this episode, Microsoft's 86 00:05:04,880 --> 00:05:09,359 Speaker 1: National Chief Technology Officer for Australia and New Zealand, Sarah Coney, 87 00:05:09,800 --> 00:05:12,800 Speaker 1: joins me to unpack what frontier firms are doing differently, 88 00:05:13,320 --> 00:05:16,479 Speaker 1: why so many KIWI companies are stuck in pilot mode, 89 00:05:16,800 --> 00:05:19,320 Speaker 1: and the concrete steps we need to take in twenty 90 00:05:19,360 --> 00:05:22,800 Speaker 1: twenty six to bridge the AI divide and turn our 91 00:05:22,839 --> 00:05:27,920 Speaker 1: application building strengths into real competitive advantage. Here's my interview 92 00:05:27,920 --> 00:05:42,159 Speaker 1: with Microsoft's Sarah Conney. Sarah Knney, Welcome to the Business 93 00:05:42,240 --> 00:05:44,159 Speaker 1: of Tech. How are you doing very well? 94 00:05:44,160 --> 00:05:44,960 Speaker 3: Thank you for having me. 95 00:05:45,240 --> 00:05:47,320 Speaker 1: Yeah, great to have you on. We're just into the 96 00:05:47,400 --> 00:05:51,120 Speaker 1: new year now, another year of what's probably going to be, 97 00:05:51,640 --> 00:05:54,960 Speaker 1: you know, a wild ride in the world of artificial intelligence. 98 00:05:55,240 --> 00:05:59,080 Speaker 1: Twenty twenty five was was massive on all sorts of levels, 99 00:05:59,120 --> 00:06:02,960 Speaker 1: from the infrastructure investment that's gone on in AI through 100 00:06:03,000 --> 00:06:06,920 Speaker 1: to the actual development of agentic AI. It was really 101 00:06:06,960 --> 00:06:09,919 Speaker 1: the year of AI agents and that's only going to 102 00:06:09,960 --> 00:06:11,800 Speaker 1: accelerate in twenty twenty six. 103 00:06:12,160 --> 00:06:14,360 Speaker 4: But just before the break, Sarah. 104 00:06:14,800 --> 00:06:19,280 Speaker 1: Microsoft commissioned IDC to do some research, so we're going 105 00:06:19,279 --> 00:06:22,160 Speaker 1: to have a look at that some really interesting results. 106 00:06:22,160 --> 00:06:25,440 Speaker 1: This was four thousand businesses across the world that were 107 00:06:26,160 --> 00:06:29,880 Speaker 1: surveyed about their uptake of AI, and it's quite good 108 00:06:29,880 --> 00:06:32,200 Speaker 1: because there was a decent cohort of Australia and New 109 00:06:32,240 --> 00:06:36,800 Speaker 1: Zealand organizations and business people surveyed there as well. I 110 00:06:36,800 --> 00:06:39,960 Speaker 1: guess the headline figure there is, you know you're really 111 00:06:40,000 --> 00:06:44,640 Speaker 1: interested in because Microsoft Commission is in those frontier firms, 112 00:06:44,640 --> 00:06:47,200 Speaker 1: who are the ones that are leading the way on 113 00:06:47,360 --> 00:06:52,680 Speaker 1: uptake of AI IDC identified around the world. Globally, twenty 114 00:06:52,680 --> 00:06:56,840 Speaker 1: two percent of firms could be considered frontier firms. Now 115 00:06:56,839 --> 00:06:59,400 Speaker 1: when you drill down to Australia and New Zealand are 116 00:06:59,480 --> 00:07:03,120 Speaker 1: part of the world, only eight percent frontier firms. To 117 00:07:03,240 --> 00:07:06,800 Speaker 1: start off, describe for us what actually a frontier firm is. 118 00:07:07,080 --> 00:07:09,359 Speaker 3: Yeah, it's an interesting concept. And you know, when we 119 00:07:09,400 --> 00:07:11,960 Speaker 3: think about frontier and you've just talked to all the 120 00:07:12,000 --> 00:07:15,280 Speaker 3: different evolutions we've seen over the last twelve months, in 121 00:07:15,320 --> 00:07:18,440 Speaker 3: the last three years in terms of AI, it's those 122 00:07:18,600 --> 00:07:21,560 Speaker 3: organizations that are looking at where can we use AI, 123 00:07:21,840 --> 00:07:24,240 Speaker 3: all the different places that we can put it in 124 00:07:24,280 --> 00:07:27,480 Speaker 3: a really intentional way. And so we often think about 125 00:07:27,520 --> 00:07:30,560 Speaker 3: frontier through three phases. You know. The first phase is 126 00:07:30,560 --> 00:07:33,240 Speaker 3: where most of us are right now, where where we 127 00:07:33,320 --> 00:07:35,920 Speaker 3: have some sort of AI assistant working alongside us, so 128 00:07:36,000 --> 00:07:40,960 Speaker 3: a human with an AI assistant. The second frontier is 129 00:07:41,160 --> 00:07:44,240 Speaker 3: where you have agents as part of that team. And 130 00:07:44,280 --> 00:07:48,000 Speaker 3: then that third frontier is your entire functions or departments 131 00:07:48,040 --> 00:07:51,320 Speaker 3: that are operating using agents with human oversight over the 132 00:07:51,320 --> 00:07:54,040 Speaker 3: top of it. And so that frontier firm is one 133 00:07:54,040 --> 00:07:56,240 Speaker 3: where they are looking at every place they can use 134 00:07:56,280 --> 00:08:00,240 Speaker 3: AI to optimize operations and move faster, get that of 135 00:08:00,240 --> 00:08:03,280 Speaker 3: customer outcomes, that whatever the focus of that organization is. 136 00:08:03,520 --> 00:08:07,520 Speaker 1: And look, we do in New Zealand have some frankly 137 00:08:07,600 --> 00:08:11,320 Speaker 1: world leading companies using AI. I'm thinking of one end 138 00:08:11,400 --> 00:08:15,200 Speaker 1: Z which I went to Salesforce their big Dreamforce conference 139 00:08:15,280 --> 00:08:17,560 Speaker 1: last year and they were held up them and Fisher 140 00:08:17,560 --> 00:08:20,680 Speaker 1: and pikel and zero as companies that are on the 141 00:08:20,720 --> 00:08:24,320 Speaker 1: forefront of rolling out agentic AI. So we do have 142 00:08:25,280 --> 00:08:28,280 Speaker 1: a handful of companies that have really picked up that 143 00:08:28,400 --> 00:08:29,720 Speaker 1: technology and run with it. 144 00:08:30,000 --> 00:08:31,520 Speaker 4: But by and large we're trailing. 145 00:08:31,560 --> 00:08:35,960 Speaker 1: Another stat from the IDC research fifty three percent of 146 00:08:36,480 --> 00:08:39,840 Speaker 1: Australian and New Zealand organizations are actually using AI today. 147 00:08:39,920 --> 00:08:43,520 Speaker 1: So that's trailing the world wide stats which is sixty 148 00:08:43,559 --> 00:08:46,440 Speaker 1: eight percent. So you know, just when you look broadly 149 00:08:46,480 --> 00:08:50,920 Speaker 1: across business, we have less businesses taking up the technology. 150 00:08:51,360 --> 00:08:54,280 Speaker 1: I guess the big question is is this a problem. 151 00:08:54,400 --> 00:08:58,040 Speaker 1: We've seen this trend and had this discussion around for instance, 152 00:08:58,160 --> 00:09:01,400 Speaker 1: cloud computing. Our firms here in New Zealand were a 153 00:09:01,400 --> 00:09:05,880 Speaker 1: bit late to migrating their applications and data to the 154 00:09:05,880 --> 00:09:08,720 Speaker 1: cloud and we were relatively late. Was great when we 155 00:09:08,760 --> 00:09:12,480 Speaker 1: got Microsoft's Hyperscale data center here in AWS, but we 156 00:09:12,480 --> 00:09:15,320 Speaker 1: were sort of late to the party on that as well. 157 00:09:16,040 --> 00:09:19,440 Speaker 1: We have low productivity in New Zealand compared to other 158 00:09:19,480 --> 00:09:22,200 Speaker 1: countries in the OECD. So a lot of people, including 159 00:09:22,400 --> 00:09:24,600 Speaker 1: people from Microsoft, have been saying there is a link 160 00:09:24,640 --> 00:09:28,480 Speaker 1: here between our slow uptake of these advanced technologies and 161 00:09:28,520 --> 00:09:32,440 Speaker 1: our progress on productivity. Yeah, keen on your insights, as 162 00:09:32,720 --> 00:09:35,439 Speaker 1: some of you've been at Microsoft ten years, Telstra before that, 163 00:09:35,480 --> 00:09:39,160 Speaker 1: and senior tech roles that link and the evidence that 164 00:09:39,200 --> 00:09:44,520 Speaker 1: sort of underpins it between slower uptake off technology and 165 00:09:44,600 --> 00:09:46,840 Speaker 1: what it means for a firm's productivity. 166 00:09:47,000 --> 00:09:50,520 Speaker 3: Actually, you just now the thing that we did another 167 00:09:50,559 --> 00:09:52,240 Speaker 3: report on some of your reports at the moment. There's 168 00:09:52,480 --> 00:09:55,640 Speaker 3: ai diffusion report that we we put out and what 169 00:09:55,679 --> 00:09:59,800 Speaker 3: it looks at is where does value set as technology 170 00:09:59,800 --> 00:10:03,400 Speaker 3: to fuses across a country, And so that's what you're 171 00:10:03,400 --> 00:10:06,360 Speaker 3: actually talking to there is, if we don't have this uptake, 172 00:10:06,400 --> 00:10:09,600 Speaker 3: if you know, New Zealand organizations don't use AI, we 173 00:10:09,640 --> 00:10:13,040 Speaker 3: will not realize the value as a nation, and therefore 174 00:10:13,320 --> 00:10:16,719 Speaker 3: we will have lower productivity, will have lower GDP when 175 00:10:16,800 --> 00:10:20,520 Speaker 3: you look at those comparative countries globally, and so we 176 00:10:20,640 --> 00:10:23,760 Speaker 3: are lagging behind. And I think what that stat actually hides, 177 00:10:23,800 --> 00:10:25,599 Speaker 3: that you know, sort of fifty three percent start or 178 00:10:25,600 --> 00:10:30,120 Speaker 3: whatever it is, actually hides is the way we're using 179 00:10:30,160 --> 00:10:32,240 Speaker 3: AI as well. So I don't know that we're using 180 00:10:32,320 --> 00:10:34,800 Speaker 3: it well like that is the other problem is that 181 00:10:34,840 --> 00:10:37,760 Speaker 3: we are using it like fancy Google is my son's 182 00:10:37,800 --> 00:10:40,760 Speaker 3: way of talking about it. We're just using it to 183 00:10:40,800 --> 00:10:44,840 Speaker 3: search for things as opposed to those higher level tasks. 184 00:10:44,880 --> 00:10:47,319 Speaker 3: And so I think when you look at this report 185 00:10:47,760 --> 00:10:51,040 Speaker 3: and global uptake, I think the actual fundamental difference I 186 00:10:51,040 --> 00:10:54,880 Speaker 3: see as well is really advanced use cases, really advanced 187 00:10:54,880 --> 00:10:59,000 Speaker 3: ways of leaning into AI globally that perhaps across Australia 188 00:10:59,040 --> 00:11:01,280 Speaker 3: and New Zealand we just aren't they. And so I 189 00:11:01,320 --> 00:11:04,920 Speaker 3: think that number actually masks how far behind we may 190 00:11:04,960 --> 00:11:08,520 Speaker 3: well be because we're not using it in sophisticated ways. 191 00:11:08,640 --> 00:11:11,920 Speaker 1: Oh absolutely, I think you know a good chunk of 192 00:11:11,920 --> 00:11:14,760 Speaker 1: that fifty three percent, and the AI Forum, for instance, 193 00:11:14,960 --> 00:11:16,720 Speaker 1: has it much higher that figure. I think they have 194 00:11:16,800 --> 00:11:19,080 Speaker 1: like eighty percent of using AI, but they don't actually 195 00:11:19,160 --> 00:11:21,600 Speaker 1: define what is use of AI. So it could be 196 00:11:21,640 --> 00:11:25,760 Speaker 1: that they've subscribed to Copilot from Microsoft and that's all 197 00:11:25,800 --> 00:11:30,840 Speaker 1: well and good, but that's really supercharging your personal productivity 198 00:11:30,840 --> 00:11:33,320 Speaker 1: when it comes to email and document management and that 199 00:11:33,360 --> 00:11:36,000 Speaker 1: sort of thing. But it's very different to agentic AI 200 00:11:36,320 --> 00:11:41,680 Speaker 1: and bundling AI into customer service and call centers and 201 00:11:41,679 --> 00:11:44,559 Speaker 1: that sort of thing. Do you see any sort of 202 00:11:44,640 --> 00:11:47,800 Speaker 1: structural reasons why New Zealand in particular has sort of 203 00:11:47,840 --> 00:11:51,160 Speaker 1: fewer frontier firms. Is it our the size of our firms, 204 00:11:51,240 --> 00:11:52,840 Speaker 1: is it our capital intensity? 205 00:11:53,000 --> 00:11:54,200 Speaker 4: Is it a skills problem? 206 00:11:54,320 --> 00:11:57,280 Speaker 3: All of those things play a role in this moment. 207 00:11:57,400 --> 00:11:59,960 Speaker 3: I think across Australia and New Zealand, we are inherently 208 00:12:00,040 --> 00:12:02,240 Speaker 3: cynical and so we ask a lot of questions and 209 00:12:02,280 --> 00:12:05,560 Speaker 3: I'm seeing that actually now almost holding us back because 210 00:12:05,720 --> 00:12:10,720 Speaker 3: organizations aren't trusting any of the systems at all, and 211 00:12:10,800 --> 00:12:13,160 Speaker 3: so they're not even leaning into try and find out 212 00:12:13,200 --> 00:12:15,400 Speaker 3: what's possible. And I think so that piece is holding 213 00:12:15,440 --> 00:12:18,480 Speaker 3: us back. That links to skills, skills and understanding. How 214 00:12:18,480 --> 00:12:21,520 Speaker 3: do we understand these systems and how they work? How 215 00:12:21,559 --> 00:12:25,040 Speaker 3: do we then build trust across this ecosystem of AI 216 00:12:25,320 --> 00:12:27,959 Speaker 3: and what that means? And so I think all those 217 00:12:27,960 --> 00:12:30,760 Speaker 3: things you talked to play a part of the role. 218 00:12:30,800 --> 00:12:33,000 Speaker 3: There is no one silverbullet. If there was, we would 219 00:12:33,000 --> 00:12:36,040 Speaker 3: have already fixed it, right. So it's a really nuanced 220 00:12:36,040 --> 00:12:40,160 Speaker 3: and challenging problem. And I think people just taking the 221 00:12:40,200 --> 00:12:43,040 Speaker 3: time to educate themselves as a really important piece of this, 222 00:12:43,280 --> 00:12:47,520 Speaker 3: not relying on their employers or their educational institutions where 223 00:12:47,520 --> 00:12:49,920 Speaker 3: they may be to teach them. They need to have 224 00:12:50,000 --> 00:12:53,679 Speaker 3: that personal curiosity. How do we create national level personal 225 00:12:53,720 --> 00:12:58,200 Speaker 3: curiosity in this moment so that everyone individually actually has 226 00:12:58,200 --> 00:13:01,800 Speaker 3: an understanding and takes the time because that curiosity is 227 00:13:01,800 --> 00:13:04,360 Speaker 3: what will serve you so powerfully as you start using 228 00:13:04,440 --> 00:13:07,360 Speaker 3: the tech as well. It always fascinates me what people 229 00:13:07,440 --> 00:13:10,079 Speaker 3: really care about in this moment, and the big question 230 00:13:10,160 --> 00:13:14,480 Speaker 3: they have is either governance, like where does my data go? 231 00:13:14,600 --> 00:13:16,439 Speaker 3: How do I know what the system is doing? How 232 00:13:16,480 --> 00:13:18,560 Speaker 3: can I trust it? Like? All of those questions are 233 00:13:18,559 --> 00:13:21,079 Speaker 3: still top of mind. But the second piece, and I 234 00:13:21,120 --> 00:13:23,240 Speaker 3: think this is where you perhaps go with some of 235 00:13:23,280 --> 00:13:26,640 Speaker 3: the questions is around that productivity piece and people's jobs. 236 00:13:26,800 --> 00:13:30,040 Speaker 3: People are really concerned about what does this actually mean, 237 00:13:30,080 --> 00:13:33,839 Speaker 3: and so that's holding people back. It's this wonderful dualism 238 00:13:33,840 --> 00:13:36,080 Speaker 3: that we have right which is everyone is using AI. 239 00:13:36,120 --> 00:13:38,640 Speaker 3: You hit a great number there, eighty percent. I think, 240 00:13:38,960 --> 00:13:41,200 Speaker 3: let's be honest, I think everyone is trying to use 241 00:13:41,240 --> 00:13:43,520 Speaker 3: it in their personal lives at least and maybe secretly 242 00:13:43,559 --> 00:13:45,719 Speaker 3: at work. We all have this feeling that it might 243 00:13:45,800 --> 00:13:49,400 Speaker 3: help us, but we're really scared that if we use 244 00:13:49,440 --> 00:13:52,520 Speaker 3: it effectively, we lose our jobs. And so this cognitive 245 00:13:52,559 --> 00:13:56,040 Speaker 3: dissonance is perhaps one of the biggest things that's holding 246 00:13:56,120 --> 00:13:59,280 Speaker 3: us back, because people don't have confidence that if they 247 00:13:59,360 --> 00:14:00,880 Speaker 3: use it and they're really at it at ed, job 248 00:14:00,880 --> 00:14:01,640 Speaker 3: will still exist. 249 00:14:01,760 --> 00:14:05,720 Speaker 1: Yeah, and look, twenty twenty five, we didn't really see 250 00:14:06,000 --> 00:14:10,360 Speaker 1: swathes of jobs going you know, to AI. What we 251 00:14:10,800 --> 00:14:13,240 Speaker 1: when I spoke to firms, and this was reflected in 252 00:14:13,360 --> 00:14:17,280 Speaker 1: AI forum research, is that a growing number of firms 253 00:14:17,400 --> 00:14:24,000 Speaker 1: are hiring less or they're leaving vacancies unfilled because they've realized, oh, 254 00:14:24,000 --> 00:14:26,600 Speaker 1: we can actually productively sort of use AI to do 255 00:14:26,640 --> 00:14:28,000 Speaker 1: some of these things. So there's a bit of a 256 00:14:28,040 --> 00:14:32,200 Speaker 1: sinking lid in some organizations and in New Zealand and 257 00:14:32,200 --> 00:14:35,240 Speaker 1: a soft economy. You know that I think is a trend. 258 00:14:35,440 --> 00:14:38,000 Speaker 1: You know a lot of firms have held off hiring, 259 00:14:38,120 --> 00:14:40,640 Speaker 1: but we haven't seen the swathes of jobs go. 260 00:14:40,680 --> 00:14:41,320 Speaker 4: What do you think in. 261 00:14:41,560 --> 00:14:45,640 Speaker 1: Twenty twenty six, particularly with agentic AI ramping up, where 262 00:14:46,200 --> 00:14:47,880 Speaker 1: you know, for instance, like one in zet has five 263 00:14:47,960 --> 00:14:50,640 Speaker 1: or six agents behind the scene doing if you want 264 00:14:50,640 --> 00:14:55,000 Speaker 1: to have a prepaid mobile accounts change, before you might 265 00:14:55,000 --> 00:14:56,440 Speaker 1: have had to go into a shop or talk to 266 00:14:56,440 --> 00:14:58,880 Speaker 1: someone on the phone, or even done it yourself on 267 00:14:58,920 --> 00:15:00,840 Speaker 1: the website. But now you can just go to a chatbot, 268 00:15:01,280 --> 00:15:04,840 Speaker 1: get the options and change. So that is actually cutting 269 00:15:04,840 --> 00:15:07,840 Speaker 1: out a number of steps in a process that will 270 00:15:07,880 --> 00:15:09,000 Speaker 1: only accelerate. 271 00:15:09,280 --> 00:15:10,960 Speaker 4: How do we deal with that? 272 00:15:11,040 --> 00:15:13,960 Speaker 1: And our employers and the workforce is sort of ready 273 00:15:14,000 --> 00:15:18,720 Speaker 1: for that acceleration in the impact of AI on labor. 274 00:15:18,880 --> 00:15:22,680 Speaker 3: So that example you've given is the perfect way of 275 00:15:22,720 --> 00:15:26,200 Speaker 3: thinking about the opportunity of AI. I think, like I think, 276 00:15:26,280 --> 00:15:30,040 Speaker 3: we all start off by getting AI assistance to help 277 00:15:30,120 --> 00:15:32,440 Speaker 3: us personally do things. But the example you've given is 278 00:15:32,440 --> 00:15:34,680 Speaker 3: where an organization has looked at an end to end 279 00:15:34,760 --> 00:15:37,400 Speaker 3: process and thought about what's the outcome. We're trying to 280 00:15:37,480 --> 00:15:42,080 Speaker 3: achieve faster service, better service, better customer outcomes, and how 281 00:15:42,080 --> 00:15:44,680 Speaker 3: can we do that differently? And they are using AI 282 00:15:44,760 --> 00:15:48,440 Speaker 3: to achieve that. That is the goal for a frontier 283 00:15:48,440 --> 00:15:51,240 Speaker 3: firm or any firm, which is how do we lift 284 00:15:51,240 --> 00:15:53,600 Speaker 3: the load from our people, which, let's be honest, so 285 00:15:53,840 --> 00:15:55,560 Speaker 3: they didn't have enough people to be able to service 286 00:15:55,560 --> 00:15:58,680 Speaker 3: all those requests anyway. So this now lifts the load 287 00:15:59,080 --> 00:16:02,680 Speaker 3: from the individuals their organization and enables them to refocus 288 00:16:02,720 --> 00:16:04,720 Speaker 3: on the really hard things, so things that AI can't 289 00:16:04,760 --> 00:16:07,320 Speaker 3: take care of, the really complex things that we love 290 00:16:07,440 --> 00:16:09,520 Speaker 3: that challenge us. And so if we come back to 291 00:16:09,520 --> 00:16:12,080 Speaker 3: the start of the question around what disruption will we 292 00:16:12,120 --> 00:16:15,000 Speaker 3: see in jobs? I think organizations that do this really 293 00:16:15,040 --> 00:16:18,320 Speaker 3: well will be looking at processes, how do they change 294 00:16:18,400 --> 00:16:22,800 Speaker 3: processes using agentic AI That then enables their teams to 295 00:16:22,960 --> 00:16:26,160 Speaker 3: pivot to different things. And I think that's really important 296 00:16:26,640 --> 00:16:30,760 Speaker 3: that they're intentful about that, because what do you do 297 00:16:30,920 --> 00:16:33,280 Speaker 3: when you have more spare time? We fill it with 298 00:16:33,360 --> 00:16:35,760 Speaker 3: rubbish like human nature, as we fill it with more 299 00:16:35,800 --> 00:16:38,720 Speaker 3: email and more teams calls. So if you have an 300 00:16:38,840 --> 00:16:42,000 Speaker 3: organization that's giving you great direction around what they want 301 00:16:42,040 --> 00:16:45,400 Speaker 3: you to now do with that time. It gives me 302 00:16:45,520 --> 00:16:50,520 Speaker 3: that safety, cognitive safety in that my job isn't being 303 00:16:50,560 --> 00:16:53,440 Speaker 3: taken away, it's being changed and changed in a really 304 00:16:53,520 --> 00:16:56,040 Speaker 3: great way, which is I'm pivoting to those strategic things 305 00:16:56,080 --> 00:16:57,840 Speaker 3: I always wish I had time for but never do, 306 00:16:58,240 --> 00:17:00,880 Speaker 3: And AI is now taking the load of things that 307 00:17:00,920 --> 00:17:03,000 Speaker 3: I just didn't really want to do. Anyway, I don't 308 00:17:03,040 --> 00:17:06,200 Speaker 3: have to triage all of my inbox to understand which 309 00:17:06,240 --> 00:17:08,760 Speaker 3: is the most important thing. AI has already been through 310 00:17:08,760 --> 00:17:11,040 Speaker 3: that and it's taken action on it. Or I don't 311 00:17:11,040 --> 00:17:12,760 Speaker 3: have to work out do I have the purchase limits 312 00:17:12,760 --> 00:17:15,080 Speaker 3: to buy a laptop for my new employee? Like AI 313 00:17:15,200 --> 00:17:17,280 Speaker 3: can do that thing for me, and I can actually 314 00:17:17,280 --> 00:17:20,520 Speaker 3: then focus on onboarding experiences for my team. So I 315 00:17:20,560 --> 00:17:24,840 Speaker 3: think if we summarize that, thinking about processes is where 316 00:17:24,840 --> 00:17:28,040 Speaker 3: the gold really sits for frontier firms, Like how do 317 00:17:28,040 --> 00:17:30,439 Speaker 3: you choose a business problem and solve that instead of 318 00:17:30,480 --> 00:17:33,760 Speaker 3: just throwing AI and everything hoping something sticks. And then 319 00:17:33,800 --> 00:17:35,720 Speaker 3: how do we help people think about their jobs in 320 00:17:35,760 --> 00:17:38,600 Speaker 3: a different way Because we all think about our jobs 321 00:17:38,640 --> 00:17:40,840 Speaker 3: and I have these ten tasks and now AI does 322 00:17:40,840 --> 00:17:42,879 Speaker 3: four of them, and that's where that feeling of AI 323 00:17:42,960 --> 00:17:45,919 Speaker 3: is taking my job comes from. But the reality is 324 00:17:46,200 --> 00:17:48,400 Speaker 3: like I now get to create the next four things 325 00:17:48,440 --> 00:17:49,879 Speaker 3: that I can do as part of that role. I 326 00:17:49,920 --> 00:17:52,760 Speaker 3: get to create the next evolution of this job. And 327 00:17:52,800 --> 00:17:55,480 Speaker 3: I see that particularly as these you know, the conversation 328 00:17:55,560 --> 00:17:58,320 Speaker 3: often goes to graduate roles. There's a lot of concern 329 00:17:58,520 --> 00:18:01,160 Speaker 3: there I be any entry level roles available. I think 330 00:18:01,160 --> 00:18:04,200 Speaker 3: it's such a fallacy, which is there'll be different entry 331 00:18:04,280 --> 00:18:07,880 Speaker 3: level roles, and those new graduates from universities and from 332 00:18:08,000 --> 00:18:10,880 Speaker 3: TAFE and college, they get to create what that new 333 00:18:11,000 --> 00:18:14,160 Speaker 3: entry entry level looks like with the assistance of AI. 334 00:18:14,320 --> 00:18:15,359 Speaker 3: What does that now become. 335 00:18:15,480 --> 00:18:18,920 Speaker 1: Yeah, and we saw just before Christmas the CEO if 336 00:18:19,320 --> 00:18:22,960 Speaker 1: Amazon said, you'd be absolutely crazy to stop hiring or 337 00:18:23,080 --> 00:18:26,800 Speaker 1: radically reduce hiring of graduates because they are the ones 338 00:18:26,840 --> 00:18:29,080 Speaker 1: that have gone deep on AI very quickly. 339 00:18:29,200 --> 00:18:31,480 Speaker 4: They're the most skilled in AI at the moment. They're 340 00:18:31,480 --> 00:18:33,560 Speaker 4: going to bring those skills into the business. 341 00:18:33,600 --> 00:18:35,840 Speaker 1: So I think that would be a mistake if we 342 00:18:35,840 --> 00:18:40,040 Speaker 1: were to do away with our hiring programs or reduce them. 343 00:18:40,359 --> 00:18:43,680 Speaker 1: But you know, the IDC research shows that those frontier 344 00:18:43,760 --> 00:18:47,400 Speaker 1: firms that you're talking about are getting roughly three times 345 00:18:47,440 --> 00:18:51,840 Speaker 1: the returns from AI compared with the slow adopters. So 346 00:18:51,960 --> 00:18:56,480 Speaker 1: that's significant. And what does that gap sort of translate 347 00:18:56,520 --> 00:18:59,679 Speaker 1: to in real terms for businesses? Are we talking here 348 00:18:59,720 --> 00:19:03,640 Speaker 1: about the frontier firms are having higher revenue growth, they're 349 00:19:03,640 --> 00:19:06,600 Speaker 1: gaining market share, they've got better margin, they're getting better 350 00:19:06,640 --> 00:19:10,120 Speaker 1: customer service. Are you actually seeing those sorts of results 351 00:19:10,119 --> 00:19:13,960 Speaker 1: when people adopt Microsoft's AI products? 352 00:19:14,000 --> 00:19:16,239 Speaker 3: Absolutely, but only if you were doing it in that 353 00:19:16,359 --> 00:19:19,320 Speaker 3: really intentional way. And so I think that's the thing 354 00:19:19,480 --> 00:19:22,360 Speaker 3: with frontier firms is you can't just accidentally become frontier. 355 00:19:22,840 --> 00:19:25,000 Speaker 3: It has to be intentional. You have to have a 356 00:19:25,080 --> 00:19:28,600 Speaker 3: planned process towards it. And the reason why they're actually 357 00:19:28,640 --> 00:19:32,199 Speaker 3: getting that kind of value is because they've chosen a 358 00:19:32,240 --> 00:19:36,080 Speaker 3: business problem. They haven't just bought AI and because AI 359 00:19:36,160 --> 00:19:38,480 Speaker 3: is cool and everyone should have AI. They've thought about it. 360 00:19:38,520 --> 00:19:41,760 Speaker 3: What is the problem we're trying to solve? Or imagine 361 00:19:41,800 --> 00:19:45,600 Speaker 3: if we could ten times the benefits in this space, 362 00:19:45,720 --> 00:19:48,320 Speaker 3: or what if we could grow our revenue by ten times? 363 00:19:48,560 --> 00:19:51,359 Speaker 3: What would that look like. They're challenging themselves to think 364 00:19:51,440 --> 00:19:55,720 Speaker 3: differently as opposed to simply adding AI to the thing 365 00:19:55,760 --> 00:19:58,560 Speaker 3: they do now. And I think that's the really important 366 00:19:58,560 --> 00:20:02,000 Speaker 3: thing is often people are great, I can build an 367 00:20:02,000 --> 00:20:04,719 Speaker 3: agent to do the thing we do today, and that's fine, 368 00:20:05,000 --> 00:20:08,000 Speaker 3: except that's just doing the same thing faster, and so 369 00:20:08,680 --> 00:20:10,840 Speaker 3: there is a point at which you cannot go any 370 00:20:10,920 --> 00:20:14,720 Speaker 3: faster doing the same thing. And that's the difference with frontier. 371 00:20:14,760 --> 00:20:18,160 Speaker 3: Frontier firms are now trying to do different things. They've 372 00:20:18,160 --> 00:20:20,919 Speaker 3: rethought what the process could be, they've rethought what the 373 00:20:21,080 --> 00:20:24,399 Speaker 3: outcome is that they're looking for, and they're using AI 374 00:20:24,520 --> 00:20:28,120 Speaker 3: to now actually fill gaps or help lift or create 375 00:20:28,560 --> 00:20:31,920 Speaker 3: seamless experiences like those examples that you had for one ends. 376 00:20:31,960 --> 00:20:34,360 Speaker 3: It like, it's a really great way of doing that thing, 377 00:20:34,520 --> 00:20:39,040 Speaker 3: completely better outcomes for everybody. But they've rethought that entire 378 00:20:39,119 --> 00:20:41,840 Speaker 3: experience versus just adding AI to the thing that is 379 00:20:41,880 --> 00:20:42,720 Speaker 3: already being done. 380 00:20:42,760 --> 00:20:45,480 Speaker 1: And only a third of organizations in Australia and New 381 00:20:45,560 --> 00:20:50,159 Speaker 1: Zealand are using AI for what IDC terms as industry 382 00:20:50,280 --> 00:20:53,879 Speaker 1: specific workflows. So you know, for instance, that one the 383 00:20:53,880 --> 00:20:56,920 Speaker 1: one ends and customer service in the telco sector, they've 384 00:20:56,960 --> 00:21:00,680 Speaker 1: got a whole process that's underpinned by agents. And I've 385 00:21:00,680 --> 00:21:03,040 Speaker 1: spoken of New Zealand firms who are doing similar things 386 00:21:03,040 --> 00:21:06,480 Speaker 1: and other industries as well. But we may be using 387 00:21:06,680 --> 00:21:11,600 Speaker 1: general use generative AI systems to maybe even to query 388 00:21:11,800 --> 00:21:15,840 Speaker 1: our own documents, our databases of information to extract information 389 00:21:16,000 --> 00:21:18,320 Speaker 1: from them, and that's helping a lot of firms, But 390 00:21:18,359 --> 00:21:22,080 Speaker 1: when it comes down to healthcare or manufacturing, we've been 391 00:21:22,600 --> 00:21:25,879 Speaker 1: a lot slower than other parts of the world and 392 00:21:25,920 --> 00:21:30,320 Speaker 1: actually applying AI to a specific industry use case. 393 00:21:30,520 --> 00:21:32,920 Speaker 3: Some of my favorite use cases are from New Zealand 394 00:21:33,000 --> 00:21:35,479 Speaker 3: and this is favorite as in all the ones I 395 00:21:35,520 --> 00:21:37,480 Speaker 3: know of and my absolutely favorite at the moment is 396 00:21:38,320 --> 00:21:42,840 Speaker 3: like the National Tale Health Service. Their ability to lean 397 00:21:42,880 --> 00:21:46,520 Speaker 3: into this space is just phenomenal to me because it's healthcare, 398 00:21:46,760 --> 00:21:50,960 Speaker 3: so you think about the privacy aspects. It's really important 399 00:21:51,320 --> 00:21:53,359 Speaker 3: you think about what they're doing, like they are serving 400 00:21:53,400 --> 00:21:56,600 Speaker 3: people in the toughest moment they're facing into. You don't 401 00:21:56,760 --> 00:21:59,320 Speaker 3: call Tallyhealth because everything's fine. You're calling because you have 402 00:21:59,359 --> 00:22:02,879 Speaker 3: an emergency of some kind. And the way they are 403 00:22:02,920 --> 00:22:05,959 Speaker 3: are leveraging AI to support their team. You know they 404 00:22:05,960 --> 00:22:10,439 Speaker 3: have a limited number of clinicians and their concern is 405 00:22:10,560 --> 00:22:12,840 Speaker 3: what about the call that we haven't yet got to? 406 00:22:13,200 --> 00:22:15,480 Speaker 3: What about the call in the queue, What about the 407 00:22:15,520 --> 00:22:19,719 Speaker 3: people who maybe in a really dark place? How do 408 00:22:19,760 --> 00:22:22,320 Speaker 3: we support them? And so they've been exploring agents to 409 00:22:22,359 --> 00:22:27,159 Speaker 3: help hold those people, to provide empathetic engagement and responses 410 00:22:27,160 --> 00:22:29,399 Speaker 3: to not clinical advice, none of those things, but just 411 00:22:29,400 --> 00:22:32,840 Speaker 3: to hold those people until a human can take them 412 00:22:33,000 --> 00:22:35,080 Speaker 3: and care for them and provide the support they need. 413 00:22:35,119 --> 00:22:38,360 Speaker 3: And I think that, to me is such a powerful 414 00:22:38,720 --> 00:22:41,120 Speaker 3: use cases, Like you're talking about industry use cases, that 415 00:22:41,160 --> 00:22:44,960 Speaker 3: to me is just gold. It is such a beautiful 416 00:22:44,960 --> 00:22:47,840 Speaker 3: way of leaning into this and finding a way of 417 00:22:48,040 --> 00:22:51,000 Speaker 3: bringing that human and AI moment together. Like it's just 418 00:22:51,400 --> 00:22:53,800 Speaker 3: it's not replacing the human care, but it's holding those 419 00:22:53,840 --> 00:22:55,880 Speaker 3: people until a human can be there for them. 420 00:22:55,920 --> 00:22:58,240 Speaker 1: I think it's brilliant and like, these are really compelling 421 00:22:58,320 --> 00:23:00,400 Speaker 1: stories and I've been writing about them for them couple 422 00:23:00,440 --> 00:23:02,679 Speaker 1: of years in business desk. But there's a sort of 423 00:23:02,680 --> 00:23:08,000 Speaker 1: a disconnect between being inspired by an understanding the potential 424 00:23:08,240 --> 00:23:13,119 Speaker 1: of these use cases to boost productivity and revenue and 425 00:23:13,119 --> 00:23:17,479 Speaker 1: customer satisfaction and our ability to actually do that. And 426 00:23:17,880 --> 00:23:19,879 Speaker 1: what do you think it primarily comes down to? Is 427 00:23:19,880 --> 00:23:22,080 Speaker 1: it a sort of an AI skills gap? Is that 428 00:23:22,160 --> 00:23:24,760 Speaker 1: the key thing. We just don't have the capability and 429 00:23:24,800 --> 00:23:28,160 Speaker 1: the confidence to roll this stuff out beyond sort of 430 00:23:28,320 --> 00:23:30,880 Speaker 1: an occasional pilot project in one part of the business. 431 00:23:31,440 --> 00:23:33,520 Speaker 3: Yeah, I think there are some key elements that I 432 00:23:33,560 --> 00:23:37,840 Speaker 3: see organizations really struggling with, and skills is a really 433 00:23:37,880 --> 00:23:41,879 Speaker 3: important piece about have we enabled the entire workforce to 434 00:23:41,880 --> 00:23:45,520 Speaker 3: step into this moment? And that's not just what is AI? 435 00:23:45,720 --> 00:23:47,240 Speaker 3: And I think there are two pieces. So one is 436 00:23:47,320 --> 00:23:49,320 Speaker 3: everyone needs to understand what AI is. We said that 437 00:23:49,359 --> 00:23:52,080 Speaker 3: at the start, but the other one is specific to 438 00:23:52,119 --> 00:23:54,600 Speaker 3: my role, like, well, how do I use this to 439 00:23:54,680 --> 00:23:56,879 Speaker 3: actually be effective in the job I'm doing? So to me, 440 00:23:56,920 --> 00:23:59,840 Speaker 3: there are two pieces that organizations need to do around 441 00:24:00,800 --> 00:24:03,359 Speaker 3: skilling for their teams, but a lot of it is 442 00:24:03,480 --> 00:24:07,679 Speaker 3: leadership as well, which is people. You need an organization 443 00:24:07,720 --> 00:24:12,199 Speaker 3: that has clear leadership. They know why they are doing this, 444 00:24:12,280 --> 00:24:13,840 Speaker 3: there is a purpose to it, like we're not just 445 00:24:13,880 --> 00:24:16,919 Speaker 3: doing it because because AI, We're doing it because it 446 00:24:16,960 --> 00:24:21,280 Speaker 3: actually serves something for the business. And so those tend 447 00:24:21,320 --> 00:24:24,439 Speaker 3: to be the two things that organizations struggle with. And 448 00:24:24,440 --> 00:24:26,199 Speaker 3: then the final piece would be I'd wrap it all 449 00:24:26,240 --> 00:24:29,320 Speaker 3: in under governance, how good is your data? How is 450 00:24:29,320 --> 00:24:33,800 Speaker 3: your security, who's responsible? What do you do when something 451 00:24:33,800 --> 00:24:36,879 Speaker 3: goes wrong? So to me, I often talk about like, 452 00:24:36,920 --> 00:24:38,960 Speaker 3: if you could do one thing in this moment, don't 453 00:24:39,119 --> 00:24:43,160 Speaker 3: do too One is that cultural skills piece and leadership, 454 00:24:43,200 --> 00:24:45,399 Speaker 3: and the other one is governance. If those are the 455 00:24:45,400 --> 00:24:48,320 Speaker 3: two things you do, you actually will be setting your 456 00:24:48,400 --> 00:24:51,720 Speaker 3: organization up for real success because those are things that 457 00:24:51,760 --> 00:24:52,600 Speaker 3: will hold you back. 458 00:24:52,800 --> 00:24:56,640 Speaker 1: And yeah, I wonder how much culture comes into play here. 459 00:24:56,720 --> 00:25:01,679 Speaker 1: Maybe we're too risk averse in Australia and New Zealand 460 00:25:01,680 --> 00:25:05,040 Speaker 1: because look, there are lots of surveys that that do 461 00:25:05,160 --> 00:25:09,520 Speaker 1: reveal that consumer trust and AI is pretty low. You know, 462 00:25:09,840 --> 00:25:12,080 Speaker 1: a lot of people have been burned by their first 463 00:25:12,720 --> 00:25:17,159 Speaker 1: experiences with chatbots. The first wave of chatbots weren't particularly effective. 464 00:25:17,880 --> 00:25:20,159 Speaker 1: That's improved a lot, So I think there, you know, 465 00:25:20,359 --> 00:25:22,720 Speaker 1: a lot of the public is coming from a point 466 00:25:22,720 --> 00:25:25,040 Speaker 1: of view if I don't know if I really trust AI, 467 00:25:25,160 --> 00:25:29,280 Speaker 1: and I'd rather deal with a human anyway. But on 468 00:25:29,320 --> 00:25:31,600 Speaker 1: the flip side of that, everyone hates sort of waiting 469 00:25:31,800 --> 00:25:34,640 Speaker 1: on a phone line for a bank or an airline 470 00:25:34,680 --> 00:25:37,840 Speaker 1: to answer their call. And if there is a number 471 00:25:37,840 --> 00:25:41,160 Speaker 1: of queries that can be handled through a chatbot interface, 472 00:25:41,680 --> 00:25:43,600 Speaker 1: it's going to save you a lot of time and frustration. 473 00:25:43,720 --> 00:25:46,199 Speaker 1: So I think people relate to that, but do we 474 00:25:46,440 --> 00:25:47,960 Speaker 1: sort of need to get a bit more on the 475 00:25:48,000 --> 00:25:50,800 Speaker 1: front foot and accept that, you know, this is a 476 00:25:50,880 --> 00:25:54,080 Speaker 1: reality now AI is a bit of a race. You know, 477 00:25:54,160 --> 00:25:56,480 Speaker 1: competitors are going to be adopting that as well. That 478 00:25:56,560 --> 00:25:59,160 Speaker 1: we need to have a higher appetite for risk. 479 00:25:59,400 --> 00:26:02,679 Speaker 3: I always coach that governance framework piece, which is you 480 00:26:02,680 --> 00:26:05,119 Speaker 3: don't want to take risks where you don't know what 481 00:26:05,160 --> 00:26:07,240 Speaker 3: the implications could be, Like, you want to understand what 482 00:26:07,280 --> 00:26:09,280 Speaker 3: the risk is that you're taking. And so an organization 483 00:26:09,359 --> 00:26:12,440 Speaker 3: should just be marching blindly into this moment. They need 484 00:26:12,480 --> 00:26:16,040 Speaker 3: to really understand what it is, the systems are capable of, 485 00:26:16,080 --> 00:26:18,240 Speaker 3: how they would operate across their data. All of those 486 00:26:18,240 --> 00:26:22,119 Speaker 3: things are really important. But great governance builds confidence and trust. 487 00:26:22,320 --> 00:26:24,840 Speaker 3: I realize how boring this sounds, and often I think 488 00:26:25,200 --> 00:26:28,000 Speaker 3: governance gets thrown in that red tape bucket. You don't 489 00:26:28,000 --> 00:26:31,640 Speaker 3: do governance, it'll slow us down. Irackon governance right now 490 00:26:31,840 --> 00:26:35,720 Speaker 3: is the biggest accelerator for innovation, like it just it 491 00:26:35,880 --> 00:26:39,400 Speaker 3: helps people feel like they can step in because they 492 00:26:39,440 --> 00:26:41,600 Speaker 3: know their organization has thought about the risks. They know 493 00:26:41,640 --> 00:26:45,040 Speaker 3: their organization has considered what might happen, and that they 494 00:26:45,160 --> 00:26:48,280 Speaker 3: feel protected and safe. And so for me, great governance 495 00:26:48,600 --> 00:26:53,000 Speaker 3: actually feeds that culture and innovation piece because you've wrapped 496 00:26:53,040 --> 00:26:55,840 Speaker 3: it all securely. It's the guardrails you need to know 497 00:26:55,920 --> 00:26:59,200 Speaker 3: that you can innovate quickly because everything has been considered 498 00:26:59,240 --> 00:27:01,040 Speaker 3: and taken care of. You have to think about a 499 00:27:01,080 --> 00:27:05,320 Speaker 3: lot of the organizations that we operate within have always 500 00:27:05,359 --> 00:27:08,080 Speaker 3: had this sense of, you know, we can't possibly get 501 00:27:08,080 --> 00:27:11,760 Speaker 3: things wrong, we can't afford to take chances, and that 502 00:27:12,480 --> 00:27:14,680 Speaker 3: culture that we have built over time is actually what's 503 00:27:14,680 --> 00:27:17,120 Speaker 3: holding us back right now because in order to use 504 00:27:17,520 --> 00:27:20,240 Speaker 3: AI really well, you have to experiment. You have to 505 00:27:20,280 --> 00:27:22,679 Speaker 3: try things. We've all had a chat with any of 506 00:27:22,720 --> 00:27:25,679 Speaker 3: your chosen chat pots and it's been terrible. And you 507 00:27:25,720 --> 00:27:28,879 Speaker 3: have to keep iterating and coaching it. Like that's not 508 00:27:28,960 --> 00:27:30,800 Speaker 3: a muscle we have because we're used to Well, I 509 00:27:30,800 --> 00:27:32,439 Speaker 3: tried it and it didn't work, So I'm going to 510 00:27:32,440 --> 00:27:35,119 Speaker 3: stop that now and try another thing. Like we have 511 00:27:35,200 --> 00:27:38,200 Speaker 3: to build a muscle of iteration and pushing and trying. 512 00:27:38,480 --> 00:27:40,480 Speaker 3: And I think that's just a little bit of a 513 00:27:40,560 --> 00:27:42,159 Speaker 3: shift that we need to try and build in as 514 00:27:42,200 --> 00:27:43,480 Speaker 3: part of that culture change. 515 00:27:43,600 --> 00:27:46,679 Speaker 1: I guess the good thing that the survey reveals is 516 00:27:46,720 --> 00:27:50,840 Speaker 1: that sixty six percent of A and Z organizations plan 517 00:27:50,920 --> 00:27:54,080 Speaker 1: to increase their AI budgets in the next twenty four months. 518 00:27:54,560 --> 00:27:57,360 Speaker 1: So that's good, although we are sort of more cautious 519 00:27:57,440 --> 00:27:59,760 Speaker 1: still compared to the rest of the world twenty seven 520 00:27:59,760 --> 00:28:03,720 Speaker 1: percent off our firms expect flat budgets around AI versus 521 00:28:03,760 --> 00:28:07,320 Speaker 1: twenty one percent, So we are lifting our spending, but 522 00:28:07,840 --> 00:28:09,760 Speaker 1: not as much as the rest of the world. So 523 00:28:10,160 --> 00:28:13,960 Speaker 1: interested in any thoughts or advice you have about how 524 00:28:14,000 --> 00:28:17,240 Speaker 1: to best allocate the dollars you do have potentially for 525 00:28:17,840 --> 00:28:21,560 Speaker 1: AI budgets. Are we typically overinvesting in some tools and 526 00:28:21,560 --> 00:28:25,120 Speaker 1: getting disappointed when we don't see the return under investing in, 527 00:28:25,440 --> 00:28:27,920 Speaker 1: for instance, getting our data house and order. You know, 528 00:28:27,920 --> 00:28:29,720 Speaker 1: where's the best bang for bucket the moment? 529 00:28:29,800 --> 00:28:32,399 Speaker 3: Yeah, And we're in such a crunched environment at the 530 00:28:32,400 --> 00:28:35,760 Speaker 3: moment that no one has spare budget to just throw 531 00:28:35,760 --> 00:28:37,320 Speaker 3: at things and hope that'll work, Which is why I 532 00:28:37,320 --> 00:28:41,000 Speaker 3: come back to that really intentional approach. Why are you 533 00:28:41,120 --> 00:28:44,600 Speaker 3: doing this? Then? If you have a really finite budget 534 00:28:44,600 --> 00:28:47,240 Speaker 3: to spend on AI, what are you going to spend 535 00:28:47,280 --> 00:28:50,200 Speaker 3: it on. And we often think about what is the 536 00:28:50,240 --> 00:28:53,680 Speaker 3: business value, like where is the real return in this moment? 537 00:28:53,800 --> 00:28:56,240 Speaker 3: And I know we've talked a lot about productivity, but 538 00:28:56,600 --> 00:28:59,520 Speaker 3: I think we use this term productivity and people have 539 00:28:59,560 --> 00:29:03,280 Speaker 3: different interpretations of what it is. When we Microsoft think 540 00:29:03,280 --> 00:29:06,960 Speaker 3: about business cases for AI right now, we look for 541 00:29:07,000 --> 00:29:10,040 Speaker 3: things that aren't just productivity, Like our return on investment 542 00:29:10,080 --> 00:29:14,520 Speaker 3: has to have something beyond productivity as a return, because 543 00:29:14,560 --> 00:29:19,240 Speaker 3: productivity is not experienced evenly across an organization, Like you 544 00:29:19,280 --> 00:29:21,320 Speaker 3: and I could have the same tool, Peter, and we 545 00:29:21,360 --> 00:29:25,160 Speaker 3: would have fundamentally different experiences of that because our individual 546 00:29:25,160 --> 00:29:28,080 Speaker 3: approach to it would be different. But if you change 547 00:29:28,080 --> 00:29:32,040 Speaker 3: a business process that is experienced evenly, like you and 548 00:29:32,120 --> 00:29:34,520 Speaker 3: I have exactly the same experience if we add AI 549 00:29:34,600 --> 00:29:37,040 Speaker 3: to a business process. So that's why we think of 550 00:29:37,240 --> 00:29:39,880 Speaker 3: Microsoft of that, what is the business viability in this 551 00:29:40,000 --> 00:29:43,680 Speaker 3: moment and finding things that go beyond productivity is really important. 552 00:29:44,440 --> 00:29:47,280 Speaker 3: But the other piece of that, it's a puzzle is desirability. 553 00:29:47,400 --> 00:29:49,600 Speaker 3: Do people actually want this thing? Because I think where 554 00:29:50,080 --> 00:29:52,800 Speaker 3: organizations end up wasting money as we build something because 555 00:29:52,800 --> 00:29:55,200 Speaker 3: it seems like a really cool idea, but no one 556 00:29:55,280 --> 00:29:58,960 Speaker 3: actually wants it, and why would anyone stop doing it 557 00:29:59,000 --> 00:30:01,080 Speaker 3: the way they've always done it? Use the new tool, 558 00:30:01,160 --> 00:30:03,080 Speaker 3: like you have to think about is there actually a 559 00:30:03,120 --> 00:30:06,560 Speaker 3: desire for this thing? And then the technology sits underneath that. 560 00:30:06,600 --> 00:30:08,960 Speaker 3: So those are the three lenses if you're trying to 561 00:30:09,000 --> 00:30:12,040 Speaker 3: find the golden place to spend your money, they are 562 00:30:12,040 --> 00:30:13,480 Speaker 3: the things that I would look at. Is there an 563 00:30:13,640 --> 00:30:17,480 Speaker 3: actual business viability beyond just productivity metrics? Is there a 564 00:30:17,520 --> 00:30:19,800 Speaker 3: desire for this single anyone actually use it? If you 565 00:30:19,840 --> 00:30:21,640 Speaker 3: build it not just because it's cool, but because it 566 00:30:21,720 --> 00:30:24,560 Speaker 3: solves the real problem and it helps people shift, and 567 00:30:24,600 --> 00:30:28,360 Speaker 3: can the technology actually help support that. Those things mean 568 00:30:28,400 --> 00:30:31,600 Speaker 3: you're not then just throwing money to the AI gods 569 00:30:31,600 --> 00:30:33,840 Speaker 3: and hoping something works like that to me, is how 570 00:30:33,880 --> 00:30:38,800 Speaker 3: you find those real solve a business problem, real business use. 571 00:30:38,720 --> 00:30:43,680 Speaker 1: Case and being really deliberate about measurability as well, actually 572 00:30:43,920 --> 00:30:46,440 Speaker 1: gauging this was the investment we put in the time 573 00:30:46,480 --> 00:30:49,080 Speaker 1: and effort, and this was the output, and can we 574 00:30:49,120 --> 00:30:52,200 Speaker 1: then scale that across the business and achieve even more 575 00:30:52,320 --> 00:30:55,920 Speaker 1: so having really good measurement and revisiting it? Is this 576 00:30:56,200 --> 00:30:58,560 Speaker 1: going on plan? Is this achieving the goals that we 577 00:30:58,600 --> 00:30:59,920 Speaker 1: set out to achieve. 578 00:31:00,160 --> 00:31:02,240 Speaker 3: Absolutely, and I think the piece that you just talk 579 00:31:02,320 --> 00:31:05,120 Speaker 3: to there a little is chain. The thing that gets 580 00:31:05,880 --> 00:31:09,840 Speaker 3: overlooked a lot is the technology piece of this is 581 00:31:09,960 --> 00:31:12,240 Speaker 3: really easy. And I know I'm a technologist saying that, 582 00:31:12,760 --> 00:31:15,800 Speaker 3: but honestly, trust me, the technology piece is the easy part. 583 00:31:15,840 --> 00:31:18,400 Speaker 3: The hard part is people, Like eighty percent of the 584 00:31:18,480 --> 00:31:21,320 Speaker 3: challenge you are going to face into is people change. 585 00:31:21,400 --> 00:31:23,880 Speaker 3: And so if you're not putting in place a proper 586 00:31:24,000 --> 00:31:27,280 Speaker 3: change program, you will not get value out of this 587 00:31:27,320 --> 00:31:32,400 Speaker 3: because you're trying to fundamentally shift how people work. And 588 00:31:33,240 --> 00:31:35,160 Speaker 3: you cannot do this, so you can't expect it to 589 00:31:35,200 --> 00:31:37,040 Speaker 3: work if you do it the way we've always done it, 590 00:31:37,040 --> 00:31:38,840 Speaker 3: like well, it's the traditional way of training people on 591 00:31:38,880 --> 00:31:42,080 Speaker 3: a new tech system. We roll out a thirty minute 592 00:31:42,120 --> 00:31:45,400 Speaker 3: video and everyone puts it on two times speed and 593 00:31:45,440 --> 00:31:47,880 Speaker 3: clicks next, next, next. We know that's what you will do. 594 00:31:48,720 --> 00:31:52,160 Speaker 3: We don't do the training in this moment, like this 595 00:31:52,200 --> 00:31:54,440 Speaker 3: is a technology that needs you to be using it 596 00:31:54,600 --> 00:31:57,200 Speaker 3: time and time again. You're trying to build a muscle. 597 00:31:57,840 --> 00:32:00,080 Speaker 3: And so if you haven't built a change program and 598 00:32:00,080 --> 00:32:02,640 Speaker 3: you haven't thought about how you move people from where 599 00:32:02,640 --> 00:32:04,480 Speaker 3: they are now, to where you need them to be. 600 00:32:05,080 --> 00:32:07,320 Speaker 3: Then all the money in the world won't actually get 601 00:32:07,320 --> 00:32:09,280 Speaker 3: you that return on investment because you could still find 602 00:32:09,280 --> 00:32:12,200 Speaker 3: the perfect use case, you could still deploy it really nicely, 603 00:32:12,240 --> 00:32:15,280 Speaker 3: but if you haven't shifted people, none of that return 604 00:32:15,360 --> 00:32:15,960 Speaker 3: comes back to you. 605 00:32:24,120 --> 00:32:28,040 Speaker 1: And just finally, Sarah, you know, IDC estimates that AI 606 00:32:28,120 --> 00:32:31,720 Speaker 1: could contribute three point seven percent of global GDP by 607 00:32:31,800 --> 00:32:34,600 Speaker 1: twenty thirty. That's sort of in line with what other 608 00:32:34,760 --> 00:32:40,040 Speaker 1: sort of analyst groups and other companies and economists have 609 00:32:40,120 --> 00:32:40,719 Speaker 1: been saying. 610 00:32:40,960 --> 00:32:41,880 Speaker 4: So that's great. 611 00:32:41,880 --> 00:32:44,400 Speaker 1: By implication, we're going to miss out on maybe some 612 00:32:44,480 --> 00:32:47,760 Speaker 1: of that if we don't have the same appetite for 613 00:32:47,960 --> 00:32:51,880 Speaker 1: uptake of AI and have as many frontier firms. What 614 00:32:51,920 --> 00:32:54,440 Speaker 1: can we do as a small sort of open economy, 615 00:32:54,800 --> 00:32:57,640 Speaker 1: like you know, New Zealand. Are there any levers you 616 00:32:57,680 --> 00:32:59,840 Speaker 1: think we should be pulling maybe as a nation at 617 00:33:00,160 --> 00:33:03,160 Speaker 1: national level to try and catch up, to try and 618 00:33:04,120 --> 00:33:08,360 Speaker 1: encourage our firms to become frontier firms when it comes 619 00:33:08,400 --> 00:33:08,760 Speaker 1: to AI. 620 00:33:08,960 --> 00:33:10,880 Speaker 3: You know what really fascinates me is that when I 621 00:33:10,920 --> 00:33:13,320 Speaker 3: look at governments across the region, I actually love the 622 00:33:13,320 --> 00:33:15,480 Speaker 3: way the New Zealand government is leaning in here like 623 00:33:15,560 --> 00:33:19,600 Speaker 3: they are leading. They've created leader boards for each of 624 00:33:19,640 --> 00:33:24,080 Speaker 3: the agencies and departments to measure how they are using AI, 625 00:33:24,240 --> 00:33:28,160 Speaker 3: which is a really powerful way of role modeling expectations. 626 00:33:29,120 --> 00:33:32,480 Speaker 3: They're putting out all sorts of capabilities across government. So 627 00:33:32,520 --> 00:33:35,640 Speaker 3: I think the environment is there, and now it's just 628 00:33:35,720 --> 00:33:38,120 Speaker 3: how do we create the appetite nationally? Like how do 629 00:33:38,200 --> 00:33:41,840 Speaker 3: we help shift the population? How do we think about diffusion? 630 00:33:41,920 --> 00:33:44,120 Speaker 3: And so for me that's like, how do you help 631 00:33:44,160 --> 00:33:47,520 Speaker 3: build national curiosity? How do we help make sure this 632 00:33:47,680 --> 00:33:50,360 Speaker 3: skilling available and that there is no one left behind, 633 00:33:50,720 --> 00:33:54,760 Speaker 3: that every element of the nation actually has an opportunity 634 00:33:54,800 --> 00:33:57,400 Speaker 3: in this moment, because that diffusion piece is where you 635 00:33:57,480 --> 00:34:00,840 Speaker 3: get that GDP uplift. Everyone understand standing and having an 636 00:34:00,840 --> 00:34:04,320 Speaker 3: opportunity results in GDP. And so for me, that's the 637 00:34:04,360 --> 00:34:06,520 Speaker 3: piece that I think perhaps there is more to be 638 00:34:06,600 --> 00:34:10,960 Speaker 3: done is skills and on national scale, and not just 639 00:34:11,040 --> 00:34:13,880 Speaker 3: for those people who already are in education, employment, in 640 00:34:13,920 --> 00:34:16,440 Speaker 3: training they will get given AI skills, Like how do 641 00:34:16,480 --> 00:34:18,759 Speaker 3: we find the people who sit outside of that? How 642 00:34:18,760 --> 00:34:22,040 Speaker 3: do we help find those groups who could get so 643 00:34:22,120 --> 00:34:25,000 Speaker 3: much more from this moment if only they had the opportunity. 644 00:34:25,200 --> 00:34:28,200 Speaker 1: Yeah, hey, and look maybe our LaGG eyed status could 645 00:34:28,239 --> 00:34:30,640 Speaker 1: be turned into an advantage. You know, when you're on 646 00:34:30,719 --> 00:34:33,920 Speaker 1: the bleeding edge or an early adopter, you can waste 647 00:34:33,920 --> 00:34:37,040 Speaker 1: a bit of money, you can make mistakes. So if 648 00:34:37,080 --> 00:34:39,800 Speaker 1: we at least learn from what some of those frontier 649 00:34:39,920 --> 00:34:43,520 Speaker 1: firms are doing around the world. Because when I talk 650 00:34:43,640 --> 00:34:46,960 Speaker 1: to companies here in academics and say, what is our 651 00:34:47,080 --> 00:34:50,600 Speaker 1: edge in AI when it comes to the application of AI, 652 00:34:51,320 --> 00:34:53,759 Speaker 1: because we're never going to build large language models, but 653 00:34:53,760 --> 00:34:57,480 Speaker 1: what can we do differently with AI? And everyone says, oh, 654 00:34:57,680 --> 00:35:01,640 Speaker 1: we'll high trust here, particularly a use of Mari data 655 00:35:01,640 --> 00:35:04,320 Speaker 1: and the work that's been done around that, and Microsoft 656 00:35:04,360 --> 00:35:07,080 Speaker 1: has been involved in that, so we're high trust. We're 657 00:35:07,120 --> 00:35:10,279 Speaker 1: really good on language and all of these and integrity 658 00:35:10,280 --> 00:35:14,040 Speaker 1: of data and sovereign data and all those sorts of things. Well, 659 00:35:14,040 --> 00:35:16,359 Speaker 1: maybe being a little bit later, but doing it really 660 00:35:16,400 --> 00:35:19,040 Speaker 1: well when we do it is actually a big advantage. 661 00:35:19,280 --> 00:35:22,040 Speaker 3: Absolutely. And you know what we think about the economics 662 00:35:22,040 --> 00:35:24,000 Speaker 3: of AI, like you think about the stack of things 663 00:35:24,080 --> 00:35:26,759 Speaker 3: you talked about, New Zealand will probably not build large 664 00:35:26,800 --> 00:35:31,160 Speaker 3: language models. The biggest economic opportunity is in applications. Building 665 00:35:31,200 --> 00:35:35,080 Speaker 3: applications and what is New Zealand great at building applications, 666 00:35:35,120 --> 00:35:37,040 Speaker 3: like you think about the innovation that comes out of 667 00:35:37,080 --> 00:35:40,200 Speaker 3: New Zealand, like this opportunity really sits here for you, 668 00:35:40,239 --> 00:35:43,120 Speaker 3: and so yes, leverage all that learning that has come 669 00:35:43,200 --> 00:35:45,200 Speaker 3: and lean into the thing that you are great at, 670 00:35:45,320 --> 00:35:48,480 Speaker 3: which is building applications. And this is a huge moment 671 00:35:48,560 --> 00:35:49,879 Speaker 3: for I think New Zealand. 672 00:35:49,520 --> 00:35:51,080 Speaker 4: To lean into great advice. 673 00:35:51,120 --> 00:35:54,840 Speaker 1: Sarah, thanks so much for talking us through that research. 674 00:35:54,920 --> 00:35:56,799 Speaker 1: We'll post links to all of that so people can 675 00:35:56,880 --> 00:36:00,719 Speaker 1: have a read and hopefully be inspired to build capability 676 00:36:00,760 --> 00:36:03,560 Speaker 1: in twenty twenty six. So thanks so much for coming 677 00:36:03,600 --> 00:36:04,399 Speaker 1: onto Business of Tech. 678 00:36:04,440 --> 00:36:05,960 Speaker 3: My pleasure, thank you so much for having me. 679 00:36:10,440 --> 00:36:11,120 Speaker 2: So there you have it. 680 00:36:11,440 --> 00:36:14,640 Speaker 1: My key takeaways from that chat with Sarah Coney about 681 00:36:14,719 --> 00:36:18,440 Speaker 1: uptake of AI and a difference between being a frontier firm. 682 00:36:18,800 --> 00:36:21,080 Speaker 2: And one that lags on AI uptake. 683 00:36:21,320 --> 00:36:25,960 Speaker 1: First thing, frontier firms treat AI as a business transformation, 684 00:36:26,160 --> 00:36:30,680 Speaker 1: not a tech bolt on. Senior leadership directly sponsor and 685 00:36:30,840 --> 00:36:35,320 Speaker 1: own the vision and the outcomes. Frontier firms start from 686 00:36:35,400 --> 00:36:38,719 Speaker 1: a specific high value use case. It might be productivity, 687 00:36:38,760 --> 00:36:41,759 Speaker 1: it could be customer experience, or it even could be 688 00:36:41,800 --> 00:36:42,520 Speaker 1: cost reduction. 689 00:36:42,680 --> 00:36:44,880 Speaker 2: They design AI around those. 690 00:36:44,760 --> 00:36:48,000 Speaker 1: Rather than playing with tools and hoping that value emerges 691 00:36:48,040 --> 00:36:53,799 Speaker 1: from the experimentation. They typically run multiple AI pilots in parallel, 692 00:36:54,480 --> 00:36:58,719 Speaker 1: learn across them, and iterate quickly, which gets them to 693 00:36:59,120 --> 00:37:02,640 Speaker 1: value faster than doing one cautious pilot at a time. 694 00:37:02,680 --> 00:37:06,360 Speaker 1: They go wide on AI quickly, and they describe themselves 695 00:37:06,400 --> 00:37:09,640 Speaker 1: as AI driven organizations. The ultimate aim is to have 696 00:37:09,760 --> 00:37:14,480 Speaker 1: AI infused across every aspect of the business. We have 697 00:37:14,640 --> 00:37:18,120 Speaker 1: less frontier firms for a complex mix of reasons. Many 698 00:37:18,280 --> 00:37:22,640 Speaker 1: organizations still frame AI as a technology project run by 699 00:37:22,719 --> 00:37:26,680 Speaker 1: it rather than a business lead change. So those sorts 700 00:37:26,680 --> 00:37:31,880 Speaker 1: of initiatives lack strong executive ownership, funding and clear links 701 00:37:31,920 --> 00:37:35,520 Speaker 1: to the core strategy of the organization. And as Sarah 702 00:37:35,520 --> 00:37:37,839 Speaker 1: pointed out, there's a real fear of job displacement. 703 00:37:38,040 --> 00:37:40,040 Speaker 2: It's a big blocker to AI. 704 00:37:40,719 --> 00:37:44,440 Speaker 1: You know, people worry about becoming visibly great at AI 705 00:37:44,760 --> 00:37:50,080 Speaker 1: could mean that they become redundant, or that AI automates 706 00:37:50,239 --> 00:37:53,239 Speaker 1: aspects of their job that they consider really important to 707 00:37:53,280 --> 00:37:57,920 Speaker 1: their identity in the organization, so that dampens down appetite 708 00:37:57,960 --> 00:38:02,680 Speaker 1: for adoption and structured GILLS programs. And there's a cultural 709 00:38:02,719 --> 00:38:05,000 Speaker 1: element as well. We are as Sarah puts it, a 710 00:38:05,040 --> 00:38:08,160 Speaker 1: cynical bunch. We see hype and run in the other direction. 711 00:38:08,840 --> 00:38:13,040 Speaker 1: But AI isn't going away. We need to embrace it, experiment, upskill, 712 00:38:13,040 --> 00:38:17,520 Speaker 1: and ultimately adopt it. The AI return on investment is 713 00:38:17,560 --> 00:38:21,000 Speaker 1: an interesting one. It's hard to achieve. Firms struggle with 714 00:38:21,280 --> 00:38:25,240 Speaker 1: ROI because they deploy tools like a copilot without deciding 715 00:38:25,320 --> 00:38:28,520 Speaker 1: up front what success actually looks like or how they'll 716 00:38:28,520 --> 00:38:32,400 Speaker 1: measure it. They roll out GENAI it might be copilot 717 00:38:32,520 --> 00:38:36,439 Speaker 1: or chat GPT enterprise, then later try to justify more 718 00:38:36,480 --> 00:38:40,720 Speaker 1: funding without baseline metrics, making it really hard to prove 719 00:38:40,840 --> 00:38:44,319 Speaker 1: the impact retrospectively. Then is it a biggie for New 720 00:38:44,400 --> 00:38:49,719 Speaker 1: Zealand organizations? Change and adoption work is too often just undercooked. 721 00:38:50,320 --> 00:38:53,520 Speaker 1: Without time for people to learn, practice and embed the 722 00:38:53,600 --> 00:38:57,520 Speaker 1: new ways of working, usage remain shallow. The organization never 723 00:38:57,600 --> 00:39:01,480 Speaker 1: realizes the productivity or innovation gain, just needed to show 724 00:39:01,680 --> 00:39:02,760 Speaker 1: a clear return. 725 00:39:03,520 --> 00:39:05,600 Speaker 2: So yeah, lots of good stuff there. 726 00:39:05,680 --> 00:39:09,000 Speaker 1: Thanks to Sarah Coney for delving into that IDC research, 727 00:39:09,160 --> 00:39:11,680 Speaker 1: which I'll put links to in the show notes. I 728 00:39:11,719 --> 00:39:14,080 Speaker 1: think she really put her finger on the issues holding 729 00:39:14,600 --> 00:39:18,600 Speaker 1: us back from the actual, tangible, true impacts that AI 730 00:39:18,760 --> 00:39:22,520 Speaker 1: can deliver. Next week, I'll talk to two outstanding key 731 00:39:22,600 --> 00:39:25,120 Speaker 1: we start up founders doing very well in the US 732 00:39:25,160 --> 00:39:28,680 Speaker 1: have raised a lot of money applying AI to robotics 733 00:39:28,920 --> 00:39:32,080 Speaker 1: with the aim of enabling faster and smarter automation and 734 00:39:32,160 --> 00:39:36,920 Speaker 1: everything from factories to driverless cars. That's next week on 735 00:39:37,000 --> 00:39:39,600 Speaker 1: the Business of Tech. I'll catch you then, have a 736 00:39:39,640 --> 00:39:40,200 Speaker 1: great week.