Revenue, Cost, or Noise?

Welcome to Applied AI Australia. I'm Ramon Rodriguez and here's the truth. Ignoring AI is probably the fastest way to fall behind. This show cuts through the noise and delivers only on what matters, growth, margins, and time. Each week, I translate AI into practical outcomes leaders can actually use. We're starting a new series called In Practice, where I sit down with Australian leaders and subject matter experts to understand how they're actually using AI to drive measurable business outcomes. Today, I'm in conversation with John Fong, the recent chief commercial officer of Domain. John spent 12 years at Google and two years at Uber in global leadership roles. Now, the framework John's going to share with us today, my big takeaway was does it meaningfully increase revenue or decrease costs? If not, it's all noise. Let's dive in. Thanks everyone and it's an honor to be here. I'd love to listen to the podcast and obviously I'd love to see uh your real world experience as well. Um John F is my name. Uh I am now transitioned actually from being uh the chief commercial officer and managing director at Domain around about half of domain. uh previously had 15 years at Google and Uber uh you know all around the world Europe uh you know London, Dublin uh America in a combination of marketing technical and sales leadership roles uh particularly as it pertains to Google cloud and some of the very first implications and uses of AI machine learning uh which we'll talk about but that grew up in Australia uh grew up in the Northshore here in Sydney uh and spent my the first part of my career uh learning about innovation systems then working at McKenzie uh before going to my MBA at Stanford so a very global career and uh very grateful to come full circle. I'll be back in Sydney with my family. Yeah. Awesome. Definitely a global career there, mate. Now, you've been at the forefront of tech for well over a decade. Some big names, Google, Uber, and most recently, Domain. What are the biggest lessons you've learned about transformation and innovation that directly apply to what we're seeing in AI in 2025? Yeah, I learned a ton of different things and I'll come to some of the frameworks that we particularly use for AI. I think the the probably the biggest thing I've learned is, you know, every great company, whether it's Australian, American, European, whatever it is, there are people in there who really desire to change the world, right? Really desire to change the world and and they're in it. Yes, there's a selfish motive about, you know, money and about learning, but people generally desire to make life better for the people they work with and better for the customers of whom they serve. And I think a lot of leadership lessons I've learned is really just unlocking that potential about helping people understand how what they do day-to-day actually makes a really big difference. And making sure the strategy of the company is well cascaded. So what is going to make the biggest impact in the market is actually what your goals are and it's actually what people are doing dayto-day. And many ways that's what I see my role is doing. Understanding what the market is offer offering us both in terms of opportunities and threats. uh figuring out what is the best strategy for the company in order to realize those opportunities and threats and then cascading that down to each one of the individuals uh on my team who are working and doing that. So, you know, most recent domain leading a team of 500 commercial sales and support people, another 200 product and tech people. Uh so the job there is really to make sure that every single person feels excited about they're working about and that's not easy. That's but that's what management leadership is all about. No, thanks for sharing that mate. That's definitely a good framework for leadership. If I just dig into your career specifically for a moment, what do you think are the biggest levers that you've pulled and you've pulled them hard in tech and transformation that's made the most impact? I I think a lot of it is actually about how customers get served. Our customers get served and this is constantly changing. You know, for example, I come mostly from the real estate industry and that has a very interesting structure because you have the CEO of say like a Ray White. You have a thousand Ray White officers each with a separate leadership and ownership and each of those officers might have five or 10 agents and each of them might have two or three assistants, right? So, it's quite an interesting I guess kind of I would say pyramid model when you've got these different stacks of leadership. Each one of those people in that leadership structure is a free agent. Yeah. Yeah, an agent go work for a different brand for an office. An office can switch alliances from Ray White to a hard to power dynamic that's shaped there. And so what we did at domain is introduce a structure where every single one of those people has someone inside domain serving them and we coordinate. That's very different to Google where you know my customers were actually uh the people who provide technology services to their own customers and they all act like individual companies and then my job is to give them a set of tools and incentives to help them build on and then Uber's different model altogether where actually we and the driver partners are serving end customers who are riding right and then our job is to help a a driver build their own business. So depending on the model of how a customer and end customer is served you need to design a interface a go to market strategy uh which is compatible with that which suits your business goals and their business goals and that's different per product it's different per industry and even the same product industry it's constantly changing with Uber for example we originally just sold Uber rides but then particularly during co we had Uber Eats we had all these different products and now Uber has like 20 30 different products they offer Every time you make a strategy change like that, you have to question is our customer model serving us and the customer and often it can get better. Yeah. Know it's a fascinating point there you make. We've all got to adapt our go to market strategy to different models, different business conditions and you know what we're facing. Obviously as a business leader we've got to be able to think on our feet and make hard decisions when time get tough. So no thanks for sharing that experience about co I definitely relate there. Now, diving a little bit deeper into AI. There's obviously a lot of noise in hype at the moment. I don't even like saying the word hype because the last time I heard hype before using chat GBT was a long time ago. It seems like it's straight from chat GBT. So, there is a lot of noise in hype and it's fair to say businesses are freaking out. I think off the top of my head, it's about nine out of 10 execs say that their number one concern is AI. Nine out of 10 of them. But one in 10 feel equipped. So I think a lot of that is to do with the new shiny thing everyone's chasing at the moment. That's why I created a PI Australia because my filter is simple, right? So if it doesn't impact growth margins or time like you would say it's noise and in my words I don't really care about it. So John, you're not an AI specialist but you are a global business leader and you have implemented digital transformation at scale. How are you approaching AI at an organizational level at the moment? Yeah, I love what you're saying there, Ron, and I love the the emphasis of the applied AI podcast. It's you trying to filter out the noise, and I'll give you my framework for how I particularly do that. The way I think about AI right now is it's very much at a stage of the hype cycle that is heightened. So, you imagine that you have this hype cycle where in the beginning, no one knows about it. And I'll talk about some of those days, early days at Google where it was just not even called AI. No one really cared about it. No one really understood it. And now, particularly since Chat GBT3 or 3.5 almost 3 years ago, you've had this awakening, this great awakening of what this could be. And that has gathered momentum. It's gotten into the mainstream. And now at the stage where everybody is expecting it to change everything. And the reality is it's going to change a lot less than we think in the next year. And it's going to change a lot more what we think in the next 5 years. But we are entering I think that trough of disillusionment part of the hype suck where people are going to be disappointed that AI cannot save you as much money or generate you much more revenue than you think it can in the short term that it's going to take more work more customization more maturing the technology of customers preference to do so. So for me, it's great that AI is known everywhere. It's had this amazing explosion moment. Uh but I think there's going to be a bit of disillusionment as people grapple with, hey, a lot of the fundamentals of business have not changed. A lot of the way we're doing business has not changed. This is the icing on the cake. This is not yet the cake. Yeah. No, I do agree with a lot of that. You know, business isn't, you know, business isn't just going to change overnight. There is a hype cycle going on at the moment. One thing that I do see pretty frequently, and it's not native just to AI, but we all think that the most expensive tech stack is going to save the day or there's going to be something as the shining night and armor that you can just buy your problems away. But it doesn't work like that. But specifically with something like AI, the most expensive tech stack with that implementation and taking your people on the journey is a quick way to lose a few million dollars, right? So it's not the best use of our investment if we haven't got everything else in in place with it. I am a huge believer that AI transformation is just as much if not more about people than it is with technology. Yet, every time I speak to a business and they say they're struggling with AI and they're not getting the output that they need, when we get it back, it's generally coming down to one cause. Are your people using it effectively? Have they adopted it? Do they understand the value and output it's going to deliver to them and their job? And the answer is normally no. So, people first 100% of the time. Yeah, I think that's right on Ron. I I think the way I think about it is all transformation is a people transformation. You know, I've done a lot of sales transformations in my life. How do we serve the customer better? Fundamentally, that starts with that starts with that ends with changing the job descriptions, the structure, and in many cases the people who are on your team. Uh we've gone through a few digital transformations whether it's an Uber or domain where we are changing our tech stack. Uh really that's a change in how we do things. The technology changes, but it's ultimately changing what people do, what roles are required. And so for me, if you can't get that people stuff right, then no changes to the best technology in the world will end up being a waste of money. Thanks for that insight, John. And if we dig into your time at Google now, what are your earliest memories at Google? What was it really like? What worked? What didn't? How did they scale adoption of their products, services throughout the organization and through to their clients? Yeah. Yeah. It's been fascinated by Google for those eight years and you see a company whose greatest strength is in some ways it greatest weakness. Its greatest strength is it develops incredibly cool stuff and attracts people who are incredibly talented and creates an environment where they can change the world. That's the real positive side. The negative is there's a lot of noise, a lot of developments which are super cool, but they don't necessarily do it or change anything, you know, and I was at Google for such a long time and I saw so many different things, whether it's Google Glass or virtuality glasses or things like that that was super cool but really had no commercial impact at all. Um, and a lot of things that eventually did just took a long long time. For me, whether it's AI or other technology, the question is always the same. Does this help me make meaningfully more revenue or does it help me meaningfully displace costs? That to me are the two true tests of whether a technology is bad, is a gimmick, is something cool versus something that will meaningfully impact my business. And therefore, as a business leader, CEO, CXO, whatever it is, I need to pay attention to. It's literally does it delta costs or does it delta revenue. So for example, let's let's say that to Google. Um I remember seeing the very first versions of AI back in the early 2010s with Google and actually implication was in YouTube and I remember we had this big TGIF presentation. And thank God it's Friday, our weekly company meeting. And they would demonstrate, hey, like we can now get a machine to adverrise videos. And what they were saying is this machine can tell if this is a cat video or this is a dog video. And my reaction was like, wow, that's super cool. And my second reaction was like, what does that mean? Does that help us make more revenue or does it help us displace costs? And in that time, the answer was no. But actually over time AI was used as the back engine to do something called classification. Right? We used to use or we still do Google a lot of manual processes to put videos into categories. Suitable for children not suitable for children. This particular category not that particular category and the machine over time the AI was able to do some of that first categorization work and that help actually displace costs a little and then a meaningful amount. So that was an idea that was something where it the cool stuff came first. It took a while to actually make that business meaningful. But but that was some of the evolution I saw early on. Um I was part of Google Cloud for about 10 years. And Google Cloud sells things such as Gmail, but the main product is it's like Amazon Web Services. You're selling storage, compute, applications to people, so they don't have to have their own data center, their own mainframes, etc. now quite common place again 10-15 years ago was uh was a big shift uh you know because that time everyone had their own data center if you had any meaningful amount of computing at all I remember we were by far the most advanced in AI and machine learning and one really interesting application was in radiology where we had trained the AI model by giving it hundreds of thousands of scans of someone's chest and and abdomen and saying great this one has cancer this one does not know how cancer. So we do that time and time and time again and we were able to actually get a rate of about 90 to 95% accuracy. That is you could show a machine a a a a a a scan and it could tell you there could tell you with 95% confidence whether there was a cancer present or not present which wasn't 100% but started to approach the levels of an actual professional uh specialist in this man. So pretty cool right? You could imagine that that could make a lot of money or that could save a lot of money. This is back in the mid2010s. But the reality is that product never took off for two major reasons. Number one, there's a big difference between 90 95% and 98 99%. Right? That is still, you know, in many cases an unacceptable level of risk. So it doesn't mean that it's not useful, but it means it's not as useful because after that review, you still need a human review. So again, how much cost have you displaced? uh not that much cost but the specialist still need to look at that scan afterwards because the high level of accuracy required. So that was one reason to mostly take off. Customers weren't blocking to it. But even for the customers that did flock to it, we had another problem which is the way that we charge for these services was based on the amount of compute required in training these models. And even though AI takes up some amount of computing technology, it's actually much much less than some of these big cloud workflows where your map services or really shifting your entire storage from on premise, you know, to offremise. So actually it didn't make much money for Google and so therefore the sales people weren't that engaged and so therefore it wasn't the help you made a lot more revenue and it wasn't helping customers to face it cost and so even though it's technology which will eventually change the world it went nowhere and it was just a really small pool of customers that made use of it. So for me that was the lesson of Google where you've always got to come back to this question. Can it increase revenue or can it decrease cost for either ourselves or our customers? If it can you've got something you can really pour into and you can adjust your people and customer strategy to suit that. If you don't you're going to be pushing a rock up hill you know until suddenly change that technology it gets better or much cheaper uh and then you can make some changes. So I think the lesson that we can all take from what you've just shared is your business is going to be still measured on if your ledger is in the red or it's black or if it's in positive or negative right so going into AI investments measuring it is so so important and John I know you are how are you ensuring that investments that you make in technology and AI can be quantified because a lot of organizations obviously forgetting to measure the frameworks that I use as a as a as a as a a corporate leader is is this. It's always be skeptic. Doesn't mean you just trust people. Everyone's trying to do the right thing. Everyone's trying to do the right thing. Always be skeptical. So treating an AI application and we had many of these in demand. Okay, if we spent this much time and energy to make this, then we could make this much money. That's the basic premise of all business cases, right? Fantastic. You then have to put your CFO hat on and go, okay, okay, what are the assumptions that underly this business case? And what gives us confidence that these assumptions are correct? In particular, the most sensitive assumptions are going to be number one, how much will an average user actually pay? And number two, realistically, how many users will actually use this? And then thirdly, how much will it actually cost us to get from where we are today to mine? Not just something that's beta launched, but something good enough and supported and updated that people would actually risk their enterprise workload. So those are the three things. Those are the three areas of business case sensitivity. How much we can make per user, how many users we can realistically get, and how much will it cost to get to market. And it is human nature to be optimistic about each of those three three things. We typically overest how much money we can make and how quickly we adopt it and we typically underestimate just how hard it is to get something to market and the total cost including management type including opportunity cost of getting something to market. That's when you tend to make decisions that really don't stack up. Put your private equity venture capitalist hat on. Put your CFO hat on and really be misely with an owner's mentality of hey this is my money. This is my life savings. How therefore would I act? Would I really back this as opposed to that money somewhere else or baking it as profit? Well, these are tough questions. No, exactly. It's obviously it is so new. There's obviously, you know, quite a few unknowns. Governance and risk is a real thing particularly about AI. So, I've got two questions for you from a board and governance perspective. How are leaders meant to govern things like AI that's as disruptive as it is if they don't completely understand it? that can be applied to any sort of digital wave uh technology adoption. What's your approach to that? Yeah, I think when it comes to governance, it's very helpful to have this skeptical framework, right? It's very very helpful. So, if you think about AI, there are so many questions about, you know, who owns the data, you know, now uh can the AI actually I would say commit a crime, but what happens to the AI? There's something wrong, who's liable, those kinds of questions. And so again, I I think the the best way that I approach it as a leader is to really pause, slow down, put your skeptical hat on and say, "What could go wrong? What could go wrong? Can we do case studies of other companies that have done this, that have rolled this out? What problems have there been? Can you think about things that are wrong with AI and what might happen if the product is hallucinating? Just going to do 20 25% of the time. You would just make stuff up. Okay, made stuff up in the consumer environment. Not great. make stuff up in an enterprise environment in a highly regular era like healthcare or finance. What happens? What happens if it's wrong with regards to for detection? So I think having that mindset of skepticism and curiosity is your best antidote against being overconfident as to how you could displace a previous, you know, non-AI enterprise workload. And I think that gets to some of the issue when it comes to board governance oversight. And I'm I'm on a board now. I've obviously spent a lot of time with boards. Boards are imperfect because they are made up of imperfect humans. So, you know, our managers are not perfect. You put together five or six managers on a board and it's even more imperfect because you have a series of imperfect people who then have imperfect and essentially dysfunctional relationships. I'm not judging boards. I'm just saying this is the reality of of human enterprise. Uh including people like you and me. It's true of all of us. And so, what that means is you have to assume that things will go wrong. And so the question is if things aren't going wrong, how would you know? If the board misses something, how would you know? What's the backs stop? And to me, that's why I get questions of who's ultimately accountable, you know, who on the board is seeing themselves as responsible as a back stop to check. Is there a dysfunctional dynamic to the board? You got to assume things are getting missed. Who's on the hook? Do the CEO feel like they're on the hook? Does the chairman feel like that? As you said before, ultimately all change is human change. technology is just an enabler or just a complicate right so for me what I've observed with boards is you have to assume that we're missing things and therefore if we are missing things how will we detect that we've missed something as early as possible that's what board governance history is today yeah know I love that now getting a bit practical John of your recent role as chief commercial officer and MD at domain you launched quite a few AI products which ones drove growth which ones actually moved from pilots to proven concepts that scaled and improved the business meaningfully. Can you walk us through a few examples and what you're proud of and what were the challenges you faced? Yeah, I'll talk about uh one or two and I'll talk about what I'm proud of and some of the lessons learned. Uh so I think one key lesson with AI is particularly as LLMs get smarter and smarter, it's actually about the data actually about the data, right? So for example, domain has some incredible data both things that are provided by the government by the value general uh but also signals that we get from people selling their house like you know the things about their house plus data that we observe about people's browsing habits and things like that and we combine that data to make predictions and if you're a real estate agent the most important thing you do is sell the house but in order to sell the house you have to get the right to sell the house and this is called a mandate. Uh, and what happens is the most important thing that a a real estate agent can do to grow their business is not just do a great job selling houses, but to find people who are trying to sell their house and get to them so they can pitch them, hey, I think you I know that you want to sell your house soon or if you want to sell your house soon, you should consider using me because here's my track record, etc., etc., so forth. And so what we found when we looked used the AI to look at our data is that if we combine hundreds of different signals, we can create profiles of people who are most likely to sell their house. And at any given year, the average Australian might sell their house every 15 to 20 years. So in any given year, a random person might have a 5 to 10% chance of selling their house. We created a model that increases that to about 33%. That is if if we could if you gave us a person right we could say with 33% confidence right that they are going to sell their house or not going to sell their house next 12 months and it might not sound amazing uh you know it's not like radiology 98 99% accuracy required to detect cancer but this is about us uh providing something that's much better than the 5% average so we can five to six times how accurate we are and that means if you're a real estate agent if you're going to door knock on doors or drop leaflets or go network with someone you knowing who is likely to sell their house versus not. Allows you much more streamline with your time. So was a product we've created called Lead Scope and it's a fantastic product. There's a few other products like it. We think ours is the best on the market and we launched that a few years ago and to me there's some great lessons here both of what went well and what didn't go well. Uh what went well is it's a product that you know agents absolutely loved. It made a lot of agents a lot richer and more successful because it did what it said because the predictions by I AI through the data were just so much. But what I wish we' done differently was make it super easy to use. Pretty easy to use. It links in with a with a with a a agent's database, you know, and and it tells them great this one will sell, this won't sell. Things like that. But what we found is if we made it even easier, people use it more. So one of the our big customers of this Ray White they actually integrated with their own technology with their own call list. So instead of them just having a database in their daily call list of who to call it'll literally rate them based on who's most likely to sell and just take out that extra bit of friction means that real estate agents who are busy and tired and exhausted are more likely to use that software and and achieve better results and have that virtuous cycle. Similarly, um it's a product that's used by a, you know, it's used by hundreds of agents. Used by hundreds of agents. It's pretty good. It makes us a few million dollars. Fantastic product from zero to us. The reality though is there are 50,000 agents, right? There are 50,000 agents. There are 10,000 agencies. And here you only have hundreds of agencies using it. So you have a market share of less than 10%. And the reality is this is a product that should be used by the majority in agents. It's such a great product. the the principle we talked about before remains true, which is you got to make great technology. You've got to be skeptical of the technology, but you've got to have something that's so much better than what we have that's going to make more revenue at a lower cost. But even if you make that, it still doesn't change the fact that you need to have a commercial people go to market model that makes it as easy for your people to sell and for your customers to use. Otherwise, it won't get the adoption that we could. And that's been uh I guess both the great thing and the challenge for us with a product like Lead Scope. No, that's a good example, John. It's a good example of, you know, opportunity and challenge if we look at it like that. Now, on a bit of a different angle, where do you think the lowhanging fruit is in organizations? Is it the administrative work that can be automated through AI? Is it breakthrough product ideas? Where do leaders need to be looking first to find the lowhanging fruit with automation or whatever it is? I'll be a little bit controversial here. I don't think there are that many lowhanging fruits that apply to all industries that AI can give. I there are some but they are quite specific to various functions and various industries. So for example in real estate uh the example I'll give is copyrightiting. So, every time you sell a house, you need someone to write some copy, right, about and when I say copy, this is a description of the house that's hopefully someone will read and go, "Oh, wow. That's really interesting, right? I I'm going to go to that openhouse inspection." So, that exists and that's great. Now, the reality is that's kind of cool, but does that mean that you can therefore have your marketing team? And right now, the answer is no, because you still need someone to check that copy. You still need someone to prompt that copy. So, is that going to save that person time? Yes. But you can't double or triple their workload just yet. Just yet. Well, same similar things in uh real estate with property management. You might have someone who's looking after a 100 rentals who's on the hook to help that help the the tenant fix their plumbing or help the landlord, you know, make sure their accounting is going well. In theory, um AI makes that a lot easier. You could have automatic responses. You could have chat bots. However, have any of those displaced the number of property managers we need at any given level? Not yet. And there are people hard at work to try and make a property manager more impactful to maybe they can handle 200 properties instead of 100. But that is industry specific functional specific work. So I guess my controversial take is the following. I have not seen enough progress to really displace out of cost at scale. I have seen the technology make people a bit more effective whether it's something simple as a saleserson having an AI write a pitch for them or create a presentation particular folks who may not speak English well or things of that nature. I have seen AI make people 5 or 10% better. I have seen some companies entirely based on AI. these unicorns is why only they have a few people working at them but again that remains a very niche very small part in the mainstream even though you have Microsoft doing 25% of their coding by people in AI they have not laid off 25% of their coders though it has meant they're slowing down how many computer engineers they're hiring so the key thing I heard you say three times was yet so I think your caveat is no one really knows where this train is going right or do you have another view yeah think of it like the following all Right? Think of the gap between no mobile phones to mobile phones. In the beginning, you have this technology which is really interesting. Wow. You can make a call without being plugged into a a cable, a wall. Amazing. Motorola came out with that with 1983. It still took 14 years to get the majority, not a majority, a small minority, but in millions of people using a Motorola. 1997 when the Motorola start became more mainstream. So those 14 years were about taking a technology that's awesome that can absorb huge amounts of information derive insights. This incredible technology that can create videos, pictures, words in a way that's customized to every single person. You have this incredible raw potential. The prizes I guess going to be to the technologists who can make that. But the real prizes is the companies that can commercialize that in a way that is okay. I can now sell twice as much of my product or I can spend half as much money in grading. And I don't think we're anywhere near those things right now. Nowhere near them. Nowhere near. But you're going to see eventually every industry disrupted by it as people work out and say, "Hey, whether it's the supply chain for how I make my widgets or how I train my sales team and help them make their sales decks to how I rejuvenate the experience of finding a home, eventually AI will make all those things better, but it's going to take 2, 3 years at the very least, likely 5 to 10 and maybe more to really disrupt the ways the patterns that we've generated last 20 or 30 years of how we get things done." I I've also often thought about you know print to digital you know mobile phones to cell phones and smartphones but I'm a firm believer that this is a huge disruption and uh it is going to impact every industry and I'm not a big fan of you know cutting people with AI or technology. Yeah. How can people do more? How can people do better and focus on the high value work? Exactly. Exactly. I think of like, you know, if it wasn't for the car, we'd all be on horses. And yes, when horses went to cars, lots of people who rode horses lost their jobs. But ultimately, they were doing something better. They end up be able to ride cars and be able to be a lot more productive and go a lot further rather than going to slow the horse. And I think where there's any transition like that, it's ultimately a good thing. It needs to be well managed. We need to have safety net to society. How do we retrain people, particularly folks who've been doing the same thing for decades? that's responsive to government and institutions and leaders such as us. But ultimately, technology is a good thing, even though it's really really scary. And AI is perhaps the the most definitely the most recent and maybe the biggest change of all. John, we're coming to the end of the podcast. And uh one thing I definitely want to find out before we close out for the execs and for those boards out there, what's your top tip that you've learned at big organizations, big tech, Google, Uber, that the listeners can apply today to the challenges we're going to be facing tomorrow? Yeah. I think two things and it comes back to the themes that I talked about today on your podcast. Number one, be curious. Be curious, right? Just like if someone brings something to you, don't judge it. Just go like, "Hey, how does this work? Wow, does that thing? How does it actually do that thing? What technology enables that? Could it also do this?" Curiosity is the bedrock of great leadership. It's the great relationships all come from curiosity. Great inventions come because someone said why not? Why can't we do it this way? And I think as a leader, particularly as a senior leader, the way to stay nimble is to always be curious. a one but the other side of curiosity to me is skepticism. It is good to be skeptical of AI both in terms of what it can and can't do and the risks and the downsides it presents and don't be not curious but also don't be unseptical and trust that oh yeah this is going to replace everything or change everything. We know it will eventually change everything but we don't know when and in what order and at what time. So skepticism when it comes to I don't know if we should invest that money right now because I can't see the line for a good business case that's the right decision and you save your dry powder for the real great innovations which you'll figure out being curious. So big executives out there be curious be skeptical that's not just for AI but that's for any technology I think. Now one more question on a little bit more of a light note. What tools are you actually using dayto-day? I know it's probably not the most high value question but I get it so much. What's your dayto-day mate? what is it? Yeah, for me um I I do use chat TBT the most. And so what I will typically do is if there's information, I will upload it to Chat TBT. I'll figure out the query, but I'll ask Chat TBT to make that query better for deep research, right? And typically they make the query a lot more specific, a lot better, and they ask me more prompts. That's stage one. Stage two is I will then ram that through Chat CBT, uh Claude Anthropic, uh and Google Gemini. I find each of them have different strengths and weaknesses and I'll typically run a deep research query in each of those for something that I really care about. Uh they'll take about 10 15 minutes. Google Gemini is the fastest. I found Claude AI to be the slowest and then lastly when I have this 20 30 40 page report I'll stick it in Google's notebook LN and they can talk podcasts and they'll listen podcast and for me that's how I develop overnight superpowers whether it's finding out a company that I'm looking to work for or do partnerships with. I use that kind of like series of stacked u you know LLM usage uh in order to make myself smart on things and and I find the podcast is very useful for me because I can listen to while I'm commuting while I'm running and then I can even have a conversation with the LLM about it after I made the query to ask for questions and hypotheticals. So that's been my generic usage. It isn't particularly complicated but I basically do that multiple times a day. You say it's simple, but if you step back and actually think about what you've achieved in that period of time, it's incredible, right? You've got the opportunity to learn anything and be anyone in a fraction of time. My view is AI, it turns generalists into specialist. And to your point about being curious and the enablement that provides, the business benefits is going to compound all the way to the bank. I know if you can be 10 to 15 minutes, I could get a little bit smart about pretty much anything. anyone or any company and that is something which was possible before in the age of Google but it took a lot more clicking a lot more random a lot more you know truthful searches and before Google was very very difficult so it is amazing how quickly we can upskill ourselves uh and I think you know to your question about being executive that's the real skill to learn how do you actually make your curiosity into a methodology and for me it's like oh I get a curious thought great I'm going to get smart about it the best way to get smart about it is multiple AI usage and then put in that package where I can listen to it while I'm doing something That's how I get smart quickly and stay curious and skeptical.