AI - Test or Invest?

Welcome to Applied AI Australia. I'm Ramon Rodriguez and each week we cut through the noise and focus only on what matters, growth, margins, and time. I translate AI into practical business outcomes Australian executives can actually use. Now, subscribe. I'm in conversation with Brad Granger, managing director, Podium. Today's episode is for Australian execs who tried AI but didn't get the best results. We're going to unpack why the test first mindset fails, the three questions you need to be asking to prevent bad AI investments, and how other operators are getting ROI fast. Let's get into it. Thanks for coming on the show, Brad. Mate, pleasure to be here and love the podcast, what you're doing. Love the investment in AI and yeah, happy to chat today. Brad, before we get into any use cases, I just want to anchor this one for the execs listening. Whether you're deciding to invest in the AI initiative or back it, what's the very first question we should be asking ourselves to avoid it becoming just another stalled pilot? I think the number one thing that we focus on here is how does it improve efficiency within the business? And that's a question that I constantly ask all of my leaders is when we have an idea or initiative, how can we utilize AI to improve the efficiency and the speed in which we can initiate that strategy? I think the question there is like efficiency versus growth or return on investment and I think you can actually have both. I think a big problem that AI can solve is essentially being able to you mentioned obviously CEOs and senior leaders that are listening to this. I think everyone agrees that to be able to have a consistent high performing team you have to have a culture where your people are bought into your vision and your strategy and they are aligned to enjoyment within their role. If they're fulfilled and enjoyed in their role, then obviously they're going to be able to be invested more, give you more, and your business is going to be naturally more successful. If I use the context of an AI agent, one of the big challenges we had when we first launched was, oh my god, this is going to take away from, you know, our our employees. They're going to fear that they're going to lose their job. I look at it slightly differently in the sense of every person, there's like there's three types of jobs within a business. There's jobs that people absolutely love doing and they come into work and they love doing it. And then there's roles within a business that there's elements of the job that you like and there's elements that you dislike. Sales being a perfect example of that. So a salesperson typically loves the thrill of the chase and they love closing deals. But pipeline management, prospecting, outbound phone calling, cold calling, all part of the requirement of the role, but not necessarily something they enjoy doing. So can AI replace the elements of the roles that your team don't like doing and improve the efficiency and at the end of the day a salesperson is skilled most at closing a deal and give them the environment to be able to do that. So one of the things that our AI employee sales AI employees done and automotive being our biggest industry that we work with here in Australia, the role of the AI employee is to that when a lead comes into that business, we want to be able to take the conversation from a customer that's very vaguely interested into converting them into a test drive. And so for a saleserson at a dealership, their role now is rather than doing all of that administration work, they can now come in every morning, all of their test drives are booked for the day, and they can actually do what they do best. Now, by virtue of that, they're spending 100% of their time on the skill set that they are most capable of doing. That's naturally going to increase your sales and increase your return on investment. So, I think you can achieve efficiency and growth together. It just depends on the context of how you're wanting to adopt AI. But internally, we obviously look at efficiency as number one. So Brad, just want to pressure test that for one moment for the execs listening. If I'm a CEO or a CFO and I'm deciding on an investment, when do you think I should look at AI purely as just efficiency? So I'm talking cost, speed, capacity. And when do you think I should look at it on the other side of the coin expecting it to drive growth? I think it depends on what like how how you're looking to implement AI and what tools you're looking to use. Agent or an AI employee, you can absolutely achieve both. But there are mediums that we use internally that are purely efficiency plays. Same principle can be applied on the CX side of the business. Right? If your business is running a high frequency of support tickets or inquiries that are coming through for service every single day, there's going to be an element of those that are low function tasks that can be probably achieved really easily with an email response back to the customer. Can AI solve that problem? And therefore, your skilled employees in your support function then focus on the more complex tasks that take a little bit time that traditionally may take you days to get back to that customer. But by freeing up that resource, are you able to get back to that customer essentially immediately because that's what they're spending 100% of the time. So again, a couple of examples on a sales side and go to market side and then obviously the CX side. Most executives have been told that AI is a long-term play. It depends on what it is and the context and situation. But Brad, based on what you're seeing, what's the realistic time frame to see value and what actually has to happen in order for that to be fulfilled? I'm going to probably work in the context of an employee, right? see immediate results if you're fully invested in knowing what you want to achieve from that AI solution. We don't work with any business that wants to test AI because if you want to test AI, there's the reality that your team or those that are responsible for implementation of that AI are not fully invested in setting it up correctly to achieve the right results. But in the instances where a business knows that AI or our solution is going to solve a critical problem for them and we set the implementation up correctly and have a robust structure to do that then the results can be immediate. So example of that is I go and get my car serviced. I've paid for the service. I get told my tires need doing. I don't want to do them right away and I say no. I'm not going to do that right now. in a car dealership or a service center, that work goes untouched. And essentially, there's no follow-up cadence to be able to go back to that customer and chase them up. They don't have the time. They don't have the resourcing. They don't have the people to physically go and call all those customers and say, "Hey, Brad, your tires need doing. It's been a month. You haven't booked that in. Would you like us to book that in?" Our service AI agent goes and does that. And so first customer in Australia that we launched this to within 2 hours had made their full months worth back off in deferred workbookings. Now that's that's revenue that literally was 65% opportunity that was just going out the door untouched. And for a dealership in Australia that can be anywhere from $380 to about $650,000 a year. So there's great opportunity where if the process is set up correctly, the business is invested in the implementation and their team buy into it, it can have immediate results. Where we see success not happen is where a business wants to test AI and they're not fully adopted. Yeah, that's a good ROI example. And I think this is where execs are getting burnt. They see a great demo, they roll it out, and 6 weeks later it's stalled. And it happens so often. From what you've seen, what usually breaks up the demo when it hits the real world? Yeah, I mean, everyone wants a piece of the AI bubble, and it's really easy to stand up a really sexy demonstration of a product that looks amazing and captures the hearts of a business to go, "Oh my god, I need that." And that's one part of the process. But what we've learned really quickly with AI and AI employees is that no customer conversation is is linear and no customer conversation is vertical in the sense that it follows a direct pathway every single time. And as soon as there becomes deviation to the status quo, you need to have an AI solution that's able to support that. So the difference with that being glorified FAQ that sits in a chatbot and is positioned as AI but essentially can only answer the questions that have been imported into it versus an AI agent that's able to show empathy to a customer when they need some support and they are having a conversation and you want to get an outcome. Yeah. No, I get that. So what I'm hearing is one proof. You've got to find customers that are really actually advocating for the product or service. Two, the road map. How is the product or AI agent going to get better not stale? Because there is a lot of options in today's age. And three, data. And I fundamentally believe this is the big missing link. And if you don't know much about data, please go back to previous episodes. We've spoken about it in depth. And if a vendor can't answer any of those clearly, what you're essentially saying is walk away. And I totally agree with that. Now, just to round this up for the audience today, there's a huge gap between new shiny toy and generative AI that actually moves the needle and it's going to impact your P&L and the pricing obviously is going to have to reflect. When businesses want to see real ROI, how do you think they need to be looking at AI investments? So, not exact pricing, but the level of commitment to get a meaningful return. Because if you ask 10 businesses, you're going to get 10 different responses because it varies all the way from $30 a month for a chatbt license to millions of dollars in implementation. Quite broad. What's your view? It's a good question. Our AI solutions vary, you know, on industry, the type of AI agent that you're wanting to employ. you've got integration costs into systems that your business is using to make sure that AI has all the data required to be able to respond to those customers effectively. So there's a depth of questions that need to be kind of uncovered to be able to provide that solution. Almost every solution is custom in that sense. I think the way that business owners should be treating AI is in the same context of the way you treat your employees. So, if you're looking at an AI agent, for example, you want an AI agent that when you employ them in 12 months time, you want them to have a deepened level of skill set that they're able to do a better job than they were when they first started. And that's how we take the approach for our employees. What you buy today is the worst you'll ever get. It's only going to get better and better and we'll continue to invest in growing that AI agent as you would a human. When you educate them, train them, invest in them, put time, they develop a skill set. You promote them, you pay them more, they're more valuable to your business, you'll continue to invest, put money into it, grow them, make them better, and in 12 months time, our goal is to make sure that that AI agent is so superior to the AI agent you buy today. And will that come with a cost attached to that? Of course, it will. But we're investing in that growth. Let's say it's $1,000 a month and you're getting a 10x or a 20x return. If that AI agent cost you $2,000 in 12 months time by virtue of its development, it's getting you a 15 or 20x return, then obviously the return on investment there is significantly greater. So I think don't look at it necessarily as a fixed cost. Look at it as an investment, a return on investment. and also forecast in your budgeting that if you're using a provider that's investing in the growth of that AI agent, naturally, it's going to develop, which means it's going to cost more as you go along, but the investment is worth it from the amount of money you're going to make back. So, what I like there, and I think most business leaders are going to relate to, you're looking at an AI investment like a general investment, right? So, you're going to have to expect ROI and you can't be too unreasonable. You've got to have those firm objectives and milestones in place. But what was very interesting, I've spoken about this, but you're the first person I've actually heard from a vendor perspective mention the point that you treat them like employees and they're interns at the start, per se, in a year. You give them promotions and they get better as they go. Now, moving on, Brad, one of the biggest fears execs had is clearly loss of control. And specifically from a board and governance perspective, one of the big ones is once AI is deployed, it's extremely hard to govern when things go wrong. So you need to get it right in the first place. And you're deploying this at scale. What do you think actually separates scalable AI from shiny tools that break the moment real pressure or problem hits? What we identified really quickly is when you're talking about an employee, no customer conversation is the same. They have their own brand guidelines, their own talk track, their own policies, their own procedures. And that AI agent needs to be able to follow those procedures and policies at all times. And so AI agent will literally follow a set pathway that gives full control to our customer to update live in real time any changes to policies and procedures. So what that means is not only do they have an AI employee at a fraction of the cost of a human employee, they have 100% control and ownership of that AI employee within the podium platform. Now why that's critical is generally speaking when something goes wrong and AI will still go wrong. It's not perfect. You have to lodge support tickets. You have to wait for that vendor to be able to fix the problem. it'll update and then over time their AI agent will get smarter and smarter and smarter and eventually get a little bit better. Being able to do that instantaneously means that that AI agent not only is able to fix the problem is built in a way that not only will it say yes Ramon I'll update that I'll update my response to that next time but also your current policy for me to in the way I'm supposed to respond to that says I should be doing this. So what I'm suggesting to you is we update our policy at the same time so that I don't make the same mistake. And so now you've got an AI agent that not only is out there responding to customers in a way that you want them to when you are giving it feedback, it's then providing suggestions to make sure it doesn't do the same thing again. That's the smartest employee you're ever going to have. Far smarter than a human and able to adapt to that change pretty quickly. So you layer that with the fact that an AI employee never goes to sleep, never takes annual leave, it's never sick, and it's doing the job of arguably a better job than a human on those core tasks that your current human employees don't like doing. You got a pretty powerful tool at your disposal. Got it. Now Brad, from a vendor's perspective, what are the key things buyers and executives should be asking before they invest in AI agents and Agentic AI? probably three questions come to mind straight off the bat. Number one thing, like ask to talk to that vendor. And if that vendor can't provide you with customers that are going to advocate for their product, that should be the biggest red flag straight off the bat. And if your customers are prepared to advocate for you and tell other customers how good the solution is, then you can have some confidence that it's going to solve similar challenges in your business. So I think that's the first one like ask what customers do you have and are we able to talk to some of those customers. The second one is understanding the model that they're using to develop the technology and their AI agents and ask the question what does the road map look like and what are you investing your time, energy and resourcing to make sure that over time this AI solution not only solves my problems now but is going to solve my solution in the future. Because I think the thing about AI is there's so many providers out there that are building solutions really really quickly and taking advantage of the AI bubble, but there's no foresight into what's coming in the future. And again, where businesses are getting cut short is if you're not adopting and implementing and the technology or AI solution that you're using isn't growing and developing with the speed in which AI is growing generally, then you're going to have a solution that's either not producing the results that you were expecting or it's becoming redundant in its own purpose because AI is advancing and that technology is not up to speed with the latest. So I think asking about roadmap, what their vision is and what their plan is is a critical one when you're looking at AI poise in particular. And then the last one which arguably is also equally as important is how does my data work with this AI solution? Because at the end of the day, an AI agent is only good as the information it's got available to it. So in the instance of AI employees, does the vendor or does the AI solution great with the systems, the other systems that my business uses to be able to have all the information at its disposal to be able to respond to the customer in the same way a human would. So I think those are the three critical questions that I would be asking before entertaining a conversation or signing up with a a vendor. So if those are the questions that we need to be asking to prevent bad investments from a vendor's perspective when AI is implemented properly, how quickly do you think we should be starting to see measurable ROI? And my view on it may be different from yours. So keen to hear your take. I think internally like definitely that's my view of when I look to implement AI into our business. We want to make sure that the objective that we're wanting to achieve is immediate. We work at a very frantic speed like most businesses do in our industry. And to be able to do that, we use AI to improve efficiency and generate better results faster. So I think that's certainly part of the decision-m process when we look to adopt any AI technology internally is that we're looking for an immediacy in response to achieve the outcome that we're wanting to achieve. So yeah, I think obviously there edge cases where maybe that might not be the case for larger scale initiatives, but certainly for the majority of things that we're implementing locally here, we are definitely looking for an immediacy of results. Yeah. No, no, thanks for sharing that. So if we talk about efficiency versus customer experience, where is the line between things being better, faster, cheaper for us as organization versus it being a poor experience for the customer? Where do businesses draw the line in terms of automation? Where should they stop? Where should they put the customer first? I think it's a great question. The answer to that is like you have to fundamentally understand your customers. And if you understand your customers, you're going to have a sense of what they're open to and what they're not open to in terms of AI versus dealing or interacting with a human. So I think again that depends on the department and approach in terms of what you're wanting to achieve. But I definitely feel there's elements where consumers or potential customers are still working through this phase of how they navigate with an AI agent in particular tools. But I think understanding your customer being in a position to know what they're prepared to work with. I think everyone in general now is accepting the fact that AI is not going away. And so I definitely don't feel a business should shy away from embracing AI technology. So I think it's just understand your customer journey, understand your customers and what they're prepared and then build a model that supports that. Yeah, that makes sense and it highlights why so many AI initiatives are failing at the moment in Australia. Businesses are too quick to automate steps instead of actually designing the customer journey end to end. Yeah, I think I think the biggest thing that's like critical right now is those that are providing AI solutions have an obligation to customers around how AI can benefit and how it can be utilized correctly cuz it is easy as a business owner to know that you need to implement AI, but you can go and spend tons of money. You mentioned at the start like you spend $30 here and $50 here. It all adds up. And if you're going and trying and testing all of these tools and AI solutions that are available and you don't get the outcome, you can end up going backwards pretty quickly and get sucked into, you know, AI doesn't work and taking a lot of money and not getting the outcome and I've still got to go and pay all my staff to go and do the job that I thought AI was going to do. So I think it's really important that providers and we take a lot of pride in doing this is we feel that we have not only a solution to offer to our customers but we want to be able to educate them on how to best utilize it and what the available tools are. We're also in a position that we're happy to provide suggestions to our customers around tools that we use internally to help improve efficiencies and things like that. So I think if you want to be in a position to talk to customers around AI, you need to know what you're talking about and then you have an obligation to really educate from there. Okay. What I'm hearing is the real risk is in AI itself. It's actually buying the tools without a clear problem owner or adoption plan. Now that completely solidifies what I've been talking about for the last 12 months. my framework ledger before launch. We need to have a clear problem in mind before we go and invest in the new shiny tool. Now, moving on, Brad, when AI fails in the real world, what do you normally see fails first? Is it the tech, the way it was bought, or how it was implemented? Because I think implementation is so underrated, not spoken about enough. What's your view? Yeah, I I I think more so we hear horror stories of where it's gone wrong. Like we talk to lots of customers every single day and it's the old cliche of like we've tried it and it doesn't work. And then when you start to ask questions around okay well tell me about that experiences. What does that look like? Generally speaking it's because they've been sold a flashy demo and they thought that you know this solution is going to solve all the problems of the world. And so like my number one takeaway or my number one piece of advice is if you're looking to implement AI technology, ask ask those three questions of a potential supplier that you're looking to work with and make sure that they have the plan in place that's actually not just going to solve the problem on a nice flashy demo, but when it becomes reality that you implement that piece of AI into your business, it's actually going to have success straight away, but there's a plan to grow it and make it even better. Yeah. Okay. Now, once the tech has been chosen, if implementation is going well and it's actually solving the problem that you've invested it in for, there's a huge amount of change management, right? I think internally like we're just we're super transparent like within our business around the ability for us to want to continue to adopt and embrace AI. Not only do we offer it as a solution, but we want to embrace it internally in the way that we prove efficiencies. And I think if you're open and honest with your people and your leadership team, you know, transparent with their teams, then your people have an understanding of the direction of the business. So I think transparency around that is really, really critical. And I don't necessarily think it's a bad thing for your teams to know that AI needs to be embraced and has the ability to make their jobs much easier as well. And there's always going to be an element of fear that will this eventually take my job or replace elements of my job. I think it's really important just as leaders to be able to have a really strong conversation with your staff around what the benefits are to them, but also the business. We've seen that we've had the success that we've had as a business off the back of being transparent with our people and getting them to embrace not be scared of the ability that AI has to make our business better. Brad, in closing, we've obviously spoken about quite a few things today and there's a lot of gold nuggets. The thing I hate is when I listen to a podcast or content and I think, you know, what the hell did I just go over? How can I actually apply it? So the key thing we talk about in applied AI is what can the listener take away from this podcast or this piece of content that they can apply within the next 48 hours to their business and it's going to better them their organization, their customers and people. What would be your big takeaway, your words of wisdom and advice, mate? I would say question I get asked a lot is like should I wait? Should I just go and build this myself within the business or should I use a supplier that's like already provide this solution? And the answer to that question is certainly don't wait. Like AI needs to be embraced and we need to be implementing it into as much of the business as we possibly can to improve efficiencies, gain growth, and all the things that this podcast talks about, right? like making sure that we're doing our jobs as leaders to make our businesses as most successful as they possibly can. But what I will say is if there's been any learnings over the last three years around how we've built our AI agents is that no conversation and no AI tooling is the same for every single business. There's an element of custom work that needs to go into it. And just when you got the right solution, there will be a curveball that's thrown its way and that could throw you back lots of time. And so my number one piece of advice is if you know you need to embrace it, you know you've got a problem to solve. The reality is there's probably a provider out there that's able to provide that solution for you. So ask the right questions of those providers. Make sure you're confident it's going to be able to execute on what you're looking to achieve and don't wait to implement. Get it going and the immediiacy and results will completely change your business. Last question, mate. On a personal front, what's your favorite AI tool? What do you use on a daily basis, mate? Look, I think like most people, chat GBT is probably the number one thing. I actually loved your last podcast with John where he mentioned converting deep research into podcast and I've started doing that which has been really effective. So, shout out to John if you're listening. Definitely took that from you from last session. But yeah, I'd say that's probably the main one. I think I probably spend most of my time embedded in tools that we're implementing in our business more so than personal use just by virtue of what I'm doing every single day. So I think probably more tool I'm more curious at the moment around tools that we can be implementing across different departments within our business more so than for personal use. But yeah, I mean I live basically on GPT. All right, thanks everyone. Brad Granger from Podium managing director. We'll catch you next time. If this was useful, make sure you subscribe. We publish regularly. One podcast, one newsletter each week. All you need to stay informed and execution ready. If you want to understand where your AI reading sits, head over to appliedustralia.com.au/schoolcard and take the 5minute assessment. Now remember, you can't control the speed of change, but you can control how quick you would gap. Thank you. My name's Ramon Rodriguez. This is Applied AI. Thanks for listening and I'll catch you soon.