Trascrizione completa di Why Your Best AI Ideas Wont Come From The Boardroom

So you're leading a workforce management company, which I work for, through a pretty significant transformation now. 00:02:14.240 --> 00:02:22.639 Um how are you seeing what's the biggest shift that we're seeing now in how organizations are doing this, incorporating AI? 00:02:23.120 --> 00:02:27.039 Yeah, I think there's a few ways businesses are really thinking about this. 00:02:27.280 --> 00:02:31.520 So um, and look, it's the topic of of the times, right? 00:02:31.680 --> 00:02:38.960 Um, I was even writing something this morning about, you know, this is the biggest probably transformation in 20 or 30 years. 00:02:39.199 --> 00:02:45.680 Um you know, organizations like ours, and we're taking the concept of being AI first, right? 00:02:45.919 --> 00:02:52.879 And um, you know, I think when we start to look at how organizations are operating, there's a few different ways that they're approaching that. 00:02:53.039 --> 00:02:58.639 So um one is purely, hey, can I just take what I've got now and add AI in over the top? 00:02:58.800 --> 00:03:15.599 Um another portion is, well, do we do something like OHI is trying to execute, where you know, we're changing the way our organization is structured as a business to really put AI in the center so we can start to think about how all of the parts of our business are interconnected and how that actually delivers value to a customer. 00:03:15.759 --> 00:03:16.000 Right. 00:03:16.159 --> 00:03:24.479 And I think really those are the key, when you're looking at what's in the market at the moment, those are really the key differences that are out there, especially in our space. 00:03:25.039 --> 00:03:35.759 Um now, previously off-air, we talked about the combination of AI and robotics as a real inflection point happening right now in manufacturing, uh construction. 00:03:35.919 --> 00:03:40.319 What's driving that and and when does it become impossible to ignore? 00:03:40.719 --> 00:03:45.199 Uh look, I I think for those industries, you're pretty close to it being impossible to ignore. 00:03:45.360 --> 00:03:57.520 Um, you know, I I know when we spoke about this offline a little bit, what we're thinking about as organizations, or what most executives are really thinking about, is how am I going to use the AI that's directly in front of me? 00:03:57.680 --> 00:03:57.919 Right. 00:03:58.159 --> 00:04:03.520 So you know, most people have access to Claude or they've got Chat GPT, right? 00:04:03.599 --> 00:04:06.879 They've got something perplexity, right, that they're working with. 00:04:07.199 --> 00:04:16.560 Um if we think ahead to what that actually looks like when robotics becomes a bigger part of the workforce, it's then the combination of robotics and AI, right? 00:04:16.639 --> 00:04:37.120 So if I'm in a manufacturing business, um and a few, probably two months ago, there was a um a business I was talking to where, you know, for a manufacturing workshop in Western Sydney already, they're looking at uh essentially hiring or deploying robots with AI built into it so they can actually start to execute the tasks, right? 00:04:37.360 --> 00:04:39.120 So these things are right in front of you. 00:04:39.279 --> 00:04:45.600 So if you're running a business like that, one of the things that I would also be thinking of is well, what does the makeup of the workforce actually look like for you? 00:04:45.680 --> 00:04:45.920 Right? 00:04:46.079 --> 00:04:54.079 Like right now, a lot of those companies who are using workforce software like ours, um, you know, people are clocking in and clocking out and they're out there working. 00:04:54.240 --> 00:05:08.160 If all of a sudden you're replacing a lot of those manufacturing workers with robotics, which is, I think, pretty inevitable at some point, what is the structure of your organization and how do you use people around those two sets of tools when they're combined together? 00:05:08.639 --> 00:05:20.079 So are we gonna just have like a couple of humanoid robots sitting around in the teams meeting, in the team meeting, uh the scrum in the morning, uh and and uh and then just one guy? 00:05:20.560 --> 00:05:22.000 Well, okay. 00:05:22.399 --> 00:05:31.040 If you think about our world, like if you dial into most meetings, most people already have a a person, an AI tool sitting there, right, recording the meeting, taking notes, doing those things. 00:05:31.199 --> 00:05:33.600 I I don't think it's actually that far away, right? 00:05:34.000 --> 00:05:42.959 Um so yes, I I think at some point you will end up with a workforce that is a combination of you know robotics and AI combined together with people, right? 00:05:43.120 --> 00:05:46.079 And I think that's just a logical way of thinking about that. 00:05:46.240 --> 00:05:55.360 Now, you know, if I'm being we're being reasonably frank about this, I think um Australian organizations are generally reasonably conservative, right? 00:05:55.600 --> 00:06:02.319 So, you know, this might be a three, five, ten year vision away, but that is actually going to come, right? 00:06:02.480 --> 00:06:11.519 So if I'm an executive thinking about this now, I would start exploring what innovation looks like in my space, what does that mean if I'm combining this technology with robotics? 00:06:11.680 --> 00:06:16.160 Then how do I actually structure my organization around that to get the best out of it, right? 00:06:16.319 --> 00:06:18.800 And I think ultimately AI is incredible. 00:06:18.959 --> 00:06:24.000 I think we all would look around as users and be like, this is great, it's great technology. 00:06:24.240 --> 00:06:34.480 Really understanding how you're going to use it, what is the use case, how does that evolve inside of my organization, what does my organization look to take look like to take advantage of this is really key. 00:06:34.560 --> 00:06:37.360 And a lot of times we're just let's do something, right? 00:06:37.920 --> 00:06:40.079 Let's dig into why this

matters, Dave. 00:06:40.160 --> 00:06:45.279 You've talked about traditional SaaS organizations and and to how they can move into an AI-first business. 00:06:45.600 --> 00:06:48.160 But what does that look like in practice? 00:06:49.040 --> 00:07:05.600 Yeah, uh for me, when we talk about an AI-first business, um, and you hear a lot of this from uh the startup world more than the traditional SaaS sort of environment, um, AI first is not just about putting AI in or above your product, right? 00:07:06.000 --> 00:07:14.240 For us, when you're thinking about an AI-first business, it's about how do we deploy AI as a centralized tool across our organization? 00:07:14.480 --> 00:07:35.519 So you know, if I take OHI, for example, how do I build a business model that has AI connecting our CRM, our product backlog, our roadmap, our support team or our customer success team, how am I actually using AI across all of these pieces to gain real-time value, right? 00:07:35.680 --> 00:07:48.000 And then deploying both in some of the simpler models, employing the intelligence and you know, the eugenic portions of AI to actually help you improve the outcome of your business, but also improve that outcome for your customer. 00:07:48.160 --> 00:07:48.399 Right. 00:07:48.480 --> 00:07:52.240 So if we think about that of um, we just did an upgrade today, right? 00:07:52.399 --> 00:07:59.040 Um, pushed out new new sets of um, you know, new pieces of technology, new functions, right? 00:07:59.279 --> 00:08:23.040 We would we want to be connecting those pieces and seeing what our customers are using, what is the feedback on those, what issues are coming back, what does that mean from a customer success perspective, and having those lit pieces linked together, instead of it just being, hey, I've got my traditional people over here, I've got customer success, I've got support, I've got these people, right, just going out individually, and hopefully we link the gap together as well. 00:08:23.680 --> 00:08:25.360 Which leads me to my next question. 00:08:25.600 --> 00:08:34.320 If you just you know get a subscription to Cord and just say to your people, oh go for it, just uh use these tools, um, is that the right way to go? 00:08:34.480 --> 00:08:37.039 And uh and if so, what happens? 00:08:37.840 --> 00:08:44.320 Uh well there's a level of experimentation I think that has to happen um in the business. 00:08:44.639 --> 00:08:48.080 Um it works for larger organizations, right? 00:08:48.399 --> 00:08:59.279 So if you're you know, uh the example that I was using, if you're a combank, you have not unlimited budget, but you've got a pretty large budget, you can start experimenting. 00:08:59.440 --> 00:09:07.200 You want to keep the licenses in-house, you don't want people roguely going off and buying their own license and doing stuff outside of your environment, but you want them to be doing this. 00:09:07.440 --> 00:09:14.799 For an organization like us, um we need to be a little

bit more focused on how we're using it, right? 00:09:15.120 --> 00:09:23.600 So um what happens is when people take on that experimentation, we see generally see an uptick, and then it's about well, what are we actually going to use this for? 00:09:23.679 --> 00:09:29.519 And most of the time, at the moment, businesses are getting to, yep, we're here, but aren't getting to what am I actually gonna do with this? 00:09:29.759 --> 00:09:41.279 I think for businesses like ours, uh, which are sort of in that smaller mid-market range, um, understanding how you're going to use it, what is the use case, what is the purpose, what is the return on investment, right? 00:09:41.519 --> 00:09:51.840 Um, somehow in the process of bringing AI into the world, we seem to, as a group of people, sort of gotten some of the fundamentals that we know we should be doing. 00:09:52.000 --> 00:09:53.440 What is my return on investment? 00:09:53.600 --> 00:09:54.879 Is there a reason for doing this? 00:09:55.039 --> 00:09:59.360 Is it actually going to make this better for our organization or our customers or our people, right? 00:09:59.600 --> 00:10:01.759 And I think we missed that point right now. 00:10:02.080 --> 00:10:04.159 Yeah, that that makes a lot of sense. 00:10:04.240 --> 00:10:15.600 Um, and I think we've talked about off-air, you know, the story about the business spending$50 million in AI tokens and then them going, oh, hang on, we need to talk about how we're using it. 00:10:16.240 --> 00:10:26.080 Yeah, ultimately, um that is one of the bigger pieces, like understanding what your cost base really looks like and and what the cost of using these tools is. 00:10:26.399 --> 00:10:29.360 I think this is ultimately, Dave, is where that ROI comes into play. 00:10:32.480 --> 00:10:36.240 Is there a purpose outside of just change of giving people access to the tool? 00:10:36.639 --> 00:10:37.120 Right? 00:10:37.360 --> 00:10:45.360 Um you want to be able to say, you know, I've got 21 or 30 use cases in my business. 00:10:45.600 --> 00:10:51.919 These are the really important ones that I can see will return value for our customers or for our people straight away. 00:10:52.080 --> 00:10:57.919 We want to put value there instead of just being like, you know, we mean this, I mean this in the nicest way. 00:10:58.000 --> 00:10:59.120 You know, I've known this. 00:10:59.200 --> 00:11:05.120 Like, if we open up a tool for marketing, like, here's a tool, do we actually know what it's gonna give us? 00:11:05.360 --> 00:11:05.679 Right? 00:11:05.919 --> 00:11:14.240 Um, and we need to be able to think about what that use case is, how that actually applies in in the real world, not just sort of go, hey Dave, have a great time experimenting with it. 00:11:14.480 --> 00:11:18.639 You will, you'll have heaps of fun, but it doesn't mean we're actually gonna get a result as an organization. 00:11:18.960 --> 00:11:29.759 And you know, interestingly, uh, one thing is that I was, you

know, building out agents and you know, trying to build a brain, if you like, in terms of building out my marketing organization with a lot of agents. 00:11:29.840 --> 00:11:49.279 And similarly, you were looking at the same things with a lot of the organization within Claude, and I'm sure that sideload AI approach can happen in many companies and having that one sort of brain going, hang on, let's all use it together and and and all learn from what's been uh learnt as a company as a whole, and then take that. 00:11:49.759 --> 00:12:06.240 Yeah, look, I I think there are clear, you know, if you are going through this this process for the first time, and you know, most people are, so are we, um, there needs to be clear structured models about how you control and govern what is happening with AI, right? 00:12:06.559 --> 00:12:33.120 Um, and I know you've got some very smart people who are lined up to talk about some of this as well, but um, you know, if I was thinking about how we work um inside of the organization and how that's governed, you know, someone like myself, um our COO, our uh chief product officer, um and then people underneath should all be combining together so we understand what you're using it, how you're using it, what we actually want to get out of it. 00:12:33.360 --> 00:12:35.200 Are we controlling that outcome as well? 00:12:35.440 --> 00:12:47.519 But um, you know, I think inside of that governance model as well, Dave, it's not just about you know making sure people are doing things, but it's also getting the business to collaborate across functions, right? 00:12:47.840 --> 00:12:56.480 And um when we sort of when I say that, what I mean is it's a great outcome to have marketing sort of doing these things on their own. 00:12:56.720 --> 00:13:03.120 But what is the value when I combine Dave's expertise as you know, leading marketing? 00:13:03.440 --> 00:13:18.879 Um if I take Dave, a customer success first person, a product person, and inputs from our customer, right, and add that into the AI mix and say, well, actually, as a team, as a fusion group of people, what outcome can you produce that's better than just an individual doing it on their own? 00:13:19.120 --> 00:13:34.879 Because ultimately I think as the as the world sort of moves forward in this AI space, um, the ability to respond rapidly to well, AI helps us respond more rapidly, but the ability to combine those people together is actually what's going to drive demand and outcome and value. 00:13:36.879 --> 00:13:39.919 So in terms of that's that that's really interesting. 00:13:40.000 --> 00:13:45.200 And just I'd like to pivot now to sort of work for software, the space that OHI is in. 00:13:45.360 --> 00:13:50.240 Um for people buying it and also the vendors actually creating it. 00:13:50.399 --> 00:13:57.679 Why what should they be actually be asking vendors, especially around AI and how it's priced in particular? 00:13:58.480 --> 00:14:00.960 Yeah, look, I I think pricing is a pretty interesting model. 00:14:01.039 --> 00:14:05.039 Like we you mentioned the$50 million spent, right, as a perfect example. 00:14:05.360 --> 00:14:16.240 Um there's lots of that there's so much information as well coming out about the um the economies of scale and whether this is actually beneficial for the cost, right? 00:14:16.320 --> 00:14:18.399 And it it I would not be the first person. 00:14:18.480 --> 00:14:22.000 I it has been all over the news for probably the last two weeks, right? 00:14:22.399 --> 00:14:26.000 I think if I were sitting on the other side of the fence, it's one of the first things I would be looking at. 00:14:26.159 --> 00:14:34.720 And so um, you know, in that context of price, most things are based on a pro price per token, right? 00:14:34.960 --> 00:14:45.759 So we know that there are already SaaS organizations who've moved from their per employee per month or per employee per seat or per seat model uh to actually it's a token consumption model, right? 00:14:45.919 --> 00:14:59.679 So if I was sitting on the other side of the fence with procurement, when you're going through your process now, and it's probably not right for me to say this, but um, it is still probably the right advice is like, well, if I was on procurement on the other side, how am I how how is the AI going to be used? 00:14:59.840 --> 00:15:01.200 How is it going to be priced? 00:15:01.279 --> 00:15:03.279 Is that on top of my current PEPA model? 00:15:03.440 --> 00:15:04.399 Are you replacing that? 00:15:04.480 --> 00:15:11.360 If it's pure per token sort of model, well, how do you know vendor how much you're going to, well we're going to use as part of this? 00:15:11.519 --> 00:15:11.679 Right? 00:15:11.840 --> 00:15:22.879 And I think ultimately most vendors out there who are doing this for the first time couldn't give you a great answer to say, oh, well, to build a roster, say, in workforce management, it's going to cost you roughly X, right? 00:15:23.120 --> 00:15:34.320 I think the other side of that is if even if we go back to the use case sort of discussion from a few questions ago, you also need to consider the value of the use case the vendor's putting forward, right? 00:15:34.559 --> 00:15:40.000 And um I was actually at a coffee with an old colleague of mine this morning, and we were talking about this. 00:15:40.159 --> 00:15:49.519 If you're a retail customer and you're building a roster and your you know store manager is out there building it and it takes them 15 minutes to do, right? 00:15:49.840 --> 00:15:56.639 But it costs you, you know, let's say$50 to$80 in token cost to run that. 00:15:57.360 --> 00:15:59.120 Just pulling a number out, right? 00:15:59.759 --> 00:16:05.360 Um does that give you getting the 15 minutes back, does that give you the advantage that you actually want? 00:16:05.600 --> 00:16:05.759 Right? 00:16:05.919 --> 00:16:06.080 Yeah. 00:16:06.320 --> 00:16:15.840 And so I think if I'm also

as a customer or as a prospect or completely over the other side of the fence, I would really be thinking about like, well, what are the use cases the vendors are bringing me? 00:16:16.080 --> 00:16:18.559 Does that use case actually make sense for my business? 00:16:18.720 --> 00:16:20.320 Um, what is the cost of that? 00:16:20.480 --> 00:16:23.120 Is it worth me actually doing that to get the right outcome? 00:16:23.279 --> 00:16:23.600 Right? 00:16:23.919 --> 00:16:32.480 Um and in the workforce management space, the example that I I was talking about this morning was like if you have a big centralized rostering team, it might make sense, right? 00:16:32.799 --> 00:16:42.000 In that case where you maybe it takes you 10 minutes to build a roster today and you're a cafe, doesn't make sense if you're getting charged a particular amount of money that's over your peppin' price, right? 00:16:42.159 --> 00:16:44.960 So there's a bit in that to really think about if you're on the other side. 00:16:45.279 --> 00:16:56.159 But if you've got a workforce of 5,000 people and you've got a couple of a rostering team of, you know, two or three and they use a few thousand tokens, that would be definitely would most likely be worth it. 00:16:56.480 --> 00:16:57.519 Yeah, it could be, absolutely. 00:16:57.759 --> 00:16:57.840 Right. 00:16:58.080 --> 00:17:05.279 But this is uh what I'm sort of encouraging people to look at is whether or not the ROI makes sense to take on. 00:17:05.519 --> 00:17:05.680 Right. 00:17:06.160 --> 00:17:15.839 And I this is the this might sound like a very harsh sort of way of saying it, but um everybody wants to do AI, right? 00:17:16.000 --> 00:17:18.319 It's fun, it's the new thing, it's new tech. 00:17:18.720 --> 00:17:26.400 All because you can do it doesn't mean that you should do it, or that the fundamentals of how you're evaluating how you use AI shouldn't be considered, right? 00:17:26.799 --> 00:17:27.200 Love it. 00:17:27.359 --> 00:17:28.400 Yeah, love it. 00:17:28.640 --> 00:17:37.440 Um and we were talking about sort of large workforces, and we off air we mentioned you know the mass layoffs that have happened because of AI. 00:17:37.920 --> 00:17:40.640 Um is that a strategic mistake? 00:17:40.799 --> 00:17:45.759 And um, I mean, there was the you know the Salesforce example where they sort of regretted it. 00:17:47.279 --> 00:17:52.480 Yeah, uh I'm talking about this from purely an OHI perspective. 00:17:52.640 --> 00:18:13.279 Um and well, and I'm you know not even an arm's length, I'm further than an arm's length away, but um I from my perspective when I look at that, I think about are you as an organization missing an opportunity to do more with the technology um instead of maintaining what you have, right? 00:18:13.519 --> 00:18:29.039 And so, you know, in the OHI context, um we have a development team deploying you know deploying AI into our product space allows us to do two to three and a half times more, right? 00:18:29.599 --> 00:18:38.720 That's m that is more than I can produce from my customers, it's it's giving me a the opportunity to respond better to the demands of our customers. 00:18:39.279 --> 00:18:46.880 If I'm starting to remove people because it's more efficient, yeah, that helps my margins, it helps my bottom line, it makes things look great. 00:18:47.279 --> 00:18:50.400 Does that actually help me get to the outcome I want at the end? 00:18:50.559 --> 00:18:50.720 Right. 00:18:50.799 --> 00:18:58.160 And I think when people are doing these mass amounts of layoffs at the moment, it's looking at, well, my margins look good, everything's okay. 00:18:58.400 --> 00:19:08.000 It's sort of short-term in my head, quarterly thinking, right, instead of this long-term strategic value of what the tool actually is and what value it can bring to your organization, right? 00:19:08.799 --> 00:19:11.039 Yeah, that that makes a lot of sense. 00:19:11.279 --> 00:19:18.160 So, Dave, what does a smart practical approach to AI workforce management actually look like? 00:19:18.400 --> 00:19:30.240 If you're a you know, if you're a leader listening to this and you're like, well, I've got a mid-size or a large business, how should I approach this in terms of my overall workforce management? 00:19:30.799 --> 00:19:38.319 So I I would be looking at if I'm if I'm an organization I wanted to put AI into practice, um, start with the use cases, right? 00:19:38.480 --> 00:19:39.279 How can I use this? 00:19:39.440 --> 00:19:41.039 Where do I see value? 00:19:41.200 --> 00:19:47.519 Um start to map it out across the value that it brings for your organization. 00:19:47.680 --> 00:19:51.119 Maybe think about what's the value it brings for customers, is it efficient? 00:19:51.279 --> 00:19:59.440 Start to really map out and use a, you know, Gartner has a very good table for this, but use a structured process to actually evaluate that. 00:19:59.680 --> 00:20:05.200 Once you've got those use cases and you can evaluate what that actually means for your business, pick one or two, right? 00:20:05.279 --> 00:20:11.200 Don't sort of go the experimentation point and be like, I've got 20 of these and 10 of them look great, right? 00:20:11.359 --> 00:20:25.839 Pick two or three that are really fat really sound that you can push through your organization to make sure that you understand the technology, you understand how it's going to work, but you know that you're going to get a return on that. 00:20:26.079 --> 00:20:28.720 And for us, that's how we're approaching it. 00:20:28.880 --> 00:20:32.799 That's how I would be approaching it if I was in generally any other type of business. 00:20:33.039 --> 00:20:35.920 You know, you will come up with lots of great ideas, right? 00:20:36.079 --> 00:20:44.640 And your team will come up with lots of great ideas, and if you start talking to your staff, they'll have probably more ideas than you know, middle management and exec will probably

come up with as well. 00:20:44.720 --> 00:20:51.759 Like you'll get some great things, so you'll use a structured process to understand what is the value of this, am I going to get the right result? 00:20:52.000 --> 00:20:54.079 What happens if it goes wrong? 00:20:54.319 --> 00:20:55.759 Pick those things and start there. 00:20:56.079 --> 00:20:59.519 Tell me about the uh Gartner table that you said was very practical. 00:20:59.680 --> 00:21:01.839 What are the steps that they would use? 00:21:02.160 --> 00:21:06.240 Uh so um it's essentially uh like a scorecard model, Dave. 00:21:06.400 --> 00:21:13.279 Um and then the scorecard model essentially says, can I, is it valuable to us as an organization? 00:21:13.440 --> 00:21:16.480 You know, this is how we've used it, is it valuable for us for us as an organization? 00:21:16.559 --> 00:21:18.240 Is it valuable to our customers? 00:21:18.319 --> 00:21:21.599 Um, is it going to improve the efficiency of how we operate as an organization? 00:21:21.759 --> 00:21:24.559 Is it going to improve the outcomes for the customer by using this? 00:21:24.799 --> 00:21:29.119 Then we evaluate it in that process and using the same Gartner framework. 00:21:29.519 --> 00:21:31.680 Will our people be resistant to the change? 00:21:31.839 --> 00:21:32.079 Right? 00:21:32.240 --> 00:21:33.680 Which I think is really important. 00:21:33.759 --> 00:21:37.440 Um we want to make sure the people in our organization are part of the journey with us. 00:21:37.680 --> 00:21:40.240 Will our customers be resistant to that change, right? 00:21:40.319 --> 00:21:40.799 As well. 00:21:40.960 --> 00:21:51.039 And then sort of combining those together as a number and saying, well, hey, in this world, these are the things that return the best value, that make us the most efficient, um, that have the least resistance to change. 00:21:51.200 --> 00:21:53.839 Let's start there as that example, right? 00:21:54.319 --> 00:21:54.880 Fantastic. 00:21:55.039 --> 00:21:59.519 Now, Dave, we play a few little games here on uh evolving workforce. 00:21:59.599 --> 00:22:02.799 Um To finish off the podcast episode. 00:22:03.200 --> 00:22:05.759 The first one's called Humans versus AI. 00:22:05.920 --> 00:22:18.000 Now, one thing where you've talked about before is keeping the human in the loop and the idea that it's not just what humans own versus AI, but where the process, uh human process, uh human in the process is, I should say. 00:22:18.240 --> 00:22:20.640 Can you uh t tell us more about that? 00:22:21.279 --> 00:22:24.640 Um particularly about where that process should sit. 00:22:24.880 --> 00:22:25.119 Yes. 00:22:25.519 --> 00:22:40.559 I think when you think there are well my smile is I would be thinking when we talk about human in the loop, the thing that we really need to think about is the level of risk that's taking in that process and where the human needs to be. 00:22:40.720 --> 00:22:46.319 So when we were talking previously, uh you

know, I was talking about rostering processes, for example. 00:22:46.480 --> 00:22:56.079 Uh if we want someone to, if we want an AI tool to go back, look at historical data, produce a roster, where in the process, as you're mapping that along, do we want the human to actually approve that? 00:22:56.160 --> 00:22:56.319 Right? 00:22:56.480 --> 00:22:58.720 Do we want it to approve it before it produces it? 00:22:58.960 --> 00:23:00.640 Do we want it to do it without? 00:23:00.720 --> 00:23:05.200 Um, or do we want the the AI to produce the roster and then we approve it? 00:23:05.359 --> 00:23:06.960 Where does that actually sit in the process? 00:23:07.119 --> 00:23:21.119 And I think for um for users of these tools and and you know, execs who are setting these things up or managers who are using them, um, that is where you would start to think about, well, where is the risk and where do I want to stop that or start that process, right? 00:23:21.359 --> 00:23:31.359 Uh and there's I right now in our world, I would say there's not a lot of spaces where you would just be like, go, let the AI do its job and push things out. 00:23:31.519 --> 00:23:38.240 You really want to think about, well, actually, if we get this wrong, where do I want to put my gate to stop it? 00:23:38.319 --> 00:23:38.640 Right? 00:23:38.799 --> 00:23:41.680 So that's where I'd be sort of considering that human in the loop function. 00:23:41.920 --> 00:23:42.160 Yeah. 00:23:42.319 --> 00:23:44.160 And perhaps you'd have several gates. 00:23:44.480 --> 00:23:46.160 You could have several gates, right? 00:23:46.240 --> 00:23:49.519 You you could have produce it and then let me review the result. 00:23:49.599 --> 00:23:56.079 And it really just depends on how um one, how complicated the process is, how risky it could possibly be in that process as well. 00:23:56.480 --> 00:23:56.640 Yeah. 00:23:56.720 --> 00:23:56.880 Yeah. 00:23:57.200 --> 00:23:59.279 All right, our next game is called Change Your Mind. 00:23:59.359 --> 00:24:09.119 So what have you changed your mind about in terms of that how fast verticals should focus on or go broad on this stuff when you when you're thinking about it? 00:24:12.640 --> 00:24:16.160 My yeah, look, I I have changed my mind um about this. 00:24:16.400 --> 00:24:23.440 So, you know, I think where AI takes us in the long run is to be more vertical focused. 00:24:23.680 --> 00:24:33.039 Um if I look at HR, payroll, workforce management, we're very good at being horizontal because that's where the market has taken us. 00:24:33.279 --> 00:24:37.599 Um we've all wanted to gain the biggest amount of market share over a period of time. 00:24:37.839 --> 00:24:45.599 The issue with that is there are complications within side of particular markets. 00:24:45.839 --> 00:25:07.519 Uh and if we use, let's use community services and disability services, for example, that has quite unique operating requirements, not just purely from an award perspective and how

do we pay people, but that flows onto how do we actually roster and provide care for people, right? 00:25:08.000 --> 00:25:17.119 The deeper you go within that vertical, probably the more robust your uh software is going to be, the more robust your customer base is going to be. 00:25:17.440 --> 00:25:24.559 I think as we roll out uh these tools, that is ultimately going to be where we end up. 00:25:24.640 --> 00:25:26.640 And it's just where I think we'll end up. 00:25:26.720 --> 00:25:34.960 I might be wrong, but um over time we've sort of focused very much on a horizontal let's chase everything and solve most vendors. 00:25:35.119 --> 00:25:39.519 But I just don't see that being valuable to a wider customer base in the future. 00:25:39.839 --> 00:25:47.279 And I suppose another way of saying that is uh are we developing our software in terms of best of breed or uh across different different. 00:25:47.920 --> 00:25:49.279 Yeah, I think it goes further than that. 00:25:49.519 --> 00:25:49.839 Yeah. 00:25:50.079 --> 00:25:56.880 Um when we talk about this within human capital management, we talk of best of breed versus all in one. 00:25:57.039 --> 00:26:01.279 All in one we talk about, but does it have payroll, HR, and WFM? 00:26:02.000 --> 00:26:17.200 You could have best of breed, that's HR, WFM, payroll, but they're trying to take on retail hospitality, aged care, disability services, construction, manufacturing, you know, basically let's take on every industry. 00:26:17.279 --> 00:26:21.359 Uh and that is traditionally how most of these businesses have operated, right? 00:26:21.839 --> 00:26:22.160 Yeah. 00:26:22.319 --> 00:26:40.720 I think as the as we focus on building out AI-first organizations and software that follows those principles, uh, we end up into a process where the vertical should be deeper than the than wide, right? 00:26:40.880 --> 00:26:55.359 And mainly because it becomes it becomes more important in my mind to be able to solve the problem completely up and down the scale than it does to be able to do it across, you know, 10 or 15 verticals. 00:26:55.599 --> 00:27:01.519 Could we be a great AI product that focuses on 15 industries? 00:27:01.920 --> 00:27:03.440 Sure, but are they all the same? 00:27:03.599 --> 00:27:06.720 They have the same problems, the same requirements, the same issues. 00:27:06.960 --> 00:27:19.839 Right now, most software providers would get 70 to 80% close to where they need to be and then say, I've configured this for you and it's great, and there's a whole bunch of processes that look after it. 00:27:20.000 --> 00:27:28.240 And that's because the software in the SaaS world uh has been built for you know 70 or 80% of your use cases. 00:27:28.480 --> 00:27:40.799 Where in an AI world, when we start to look at what the tools can actually do, does it make sense to be more vertical to establish yourself as an organization than it does to be more horizontal? 00:27:41.359 --> 00:27:41.599 Yeah. 00:27:41.839 --> 00:27:42.480 Does that make sense? 00:27:42.720 --> 00:27:43.839 Yeah, that makes a lot of sense. 00:27:44.000 --> 00:27:46.240 Um yeah, that's that's cleared it up for me. 00:27:46.319 --> 00:27:57.039 And I in some ways that really rings true as well in terms of um, well, let's really complete one particular niche and go, yeah, we we can own this. 00:27:57.200 --> 00:27:58.240 Yeah, we've got this. 00:27:58.400 --> 00:27:58.799 Yeah. 00:27:59.680 --> 00:28:03.279 All right, our next game is um one prediction and one warning. 00:28:03.440 --> 00:28:10.319 What's the prediction that you'll make about AI in workforce management over the next, say, two years, three years? 00:28:14.720 --> 00:28:15.920 Good question. 00:28:16.160 --> 00:28:23.920 Uh I think the one prediction I would say is I I I probably followed the same line as what I just said. 00:28:24.079 --> 00:28:28.720 I I think you'll see more niche software appearing. 00:28:28.880 --> 00:28:29.200 Yeah. 00:28:29.279 --> 00:28:35.200 Uh because it becomes harder for everyone to be everything to everyone, right? 00:28:35.440 --> 00:28:39.440 Um, I think that's where we'll we'll end up in the workforce management software space. 00:28:39.680 --> 00:28:50.640 Um I think the the warning that uh I would consider uh it's probably an interesting one. 00:28:50.799 --> 00:28:55.279 Uh it could be for just AI or just workforce, you know. 00:28:56.240 --> 00:29:02.640 No, I I think the warning probably it goes back to what it's less workforce management. 00:29:02.799 --> 00:29:21.519 I think ultimately the message that I I would probably leave with as warning is similar to where we we've been talking earlier on of really rethinking what your what your business looks like as AI-centric is key for where the new world is going. 00:29:21.680 --> 00:29:28.400 And that is as valuable for a manufacturing and construction business as it is right now for a retail and hospitality business. 00:29:28.559 --> 00:29:38.160 Uh, if you know, if you're a retail and hospitality business, or retail particularly, AI and personalization uh will have such an impact if you're manufacturing. 00:29:38.240 --> 00:29:40.160 We spoke about robotics and AI. 00:29:40.400 --> 00:30:03.920 I would start to be thinking about what does the structure of my organization really look like three to five years down the track, uh, because we're very good as organizations of just sort of following a we will be okay, the technology is far down the road, that technology will be here in 18 months to two years, and in some places, some cases is already here now, but we're not using it, right? 00:30:04.000 --> 00:30:17.440 And so my warning is to be rethinking your organization and your business structure, but also to be spending time planning for that now so you know what it needs to be when you're ready or when it's there. 00:30:18.000 --> 00:30:34.400 Um and just to follow up on you know, get get your people to use the

technology now, what's a practical tip you could give to leaders to say, hey um guys, we need to start, you know, experimenting, we need to start putting this into our everyday. 00:30:34.559 --> 00:30:38.079 How would you what would you say to advise leaders? 00:30:38.160 --> 00:30:46.640 Um look, I know you and I discuss it all the time, and you've been very encouraging for everyone in our organization to use it as much as possible. 00:30:46.799 --> 00:30:53.519 Um we've all got clawed licenses and uh it's it's been very helpful and I've I've been using it extensively. 00:30:53.599 --> 00:30:55.759 Um what what would your tips be? 00:30:58.240 --> 00:31:06.319 For an organization like ours, it's probably a little bit easier because we're uh we're not as large as a you know, we don't have a thousand employees, for example. 00:31:06.559 --> 00:31:12.079 Uh what I would be encouraging is to really think about what the change process looks like. 00:31:12.240 --> 00:31:27.039 Um, keeping the same principles, you know, we've sort of touched on about governance and use cases, but encouraging your people to also proactively engage in those discussions and as part of that sort of fusion team and governance structure, right? 00:31:27.440 --> 00:31:36.799 Um ultimately I I think you'll find, you know, I I I spend I'd say 60% of my time talking about AI in our business to people. 00:31:37.039 --> 00:31:51.200 Uh but a lot of the best ideas about how to actually use the tools well come from the people who are at the front line who are actually talking to customers and want it and need to be able to improve the efficiency or the outcomes, right? 00:31:51.440 --> 00:32:00.240 So ultimately I would be encouraging them to take on the tools and experiment and be as part of that governance structure and fusion team structure, right? 00:32:00.400 --> 00:32:02.400 And I I think for us it's working, right? 00:32:02.480 --> 00:32:04.000 We're heading in the right direction. 00:32:04.160 --> 00:32:10.319 Um, it's never going to be perfect, but I think ultimately that helps you from a change perspective uh as well. 00:32:10.480 --> 00:32:12.640 So well, thanks very much, Dave. 00:32:12.720 --> 00:32:13.359 This has been great. 00:32:13.440 --> 00:32:16.319 It's not uh it's not often you get to interview your own boss. 00:32:16.480 --> 00:32:20.799 Um, yeah, so no, I've enjoyed our chat, mate, and uh yeah, it's very exciting. 00:32:21.200 --> 00:32:22.720 Anytime sounds good, thanks, Dave. 00:32:23.039 --> 00:32:23.759 So yeah, bye. 00:32:46.160 --> 00:32:47.920 Delivering new certainty. 00:32:48.160 --> 00:32:51.119 This podcast was recorded on the land of the Gadigal people. 00:32:51.359 --> 00:32:55.119 If you enjoyed this podcast, why not send to a friend or a colleague? 00:32:55.359 --> 00:33:02.960 And remember to hit follow on your podcast player to get every episode as they drop every Wednesday. 00:33:03.359 --> 00:33:04.240 Thanks for listening. 00:33:04.319 --> 00:33:07.759 I've been Dave Kenyon, and this is Evolving Workforce.