In this episode of the Trusted Adviser Podcast, Rob Pyne sits down with Andrew Gardner, Founder & Managing Director of RetireMap, and Ivon Gower, Operations Manager of Retiremap, to explore how artificial intelligence, deterministic modelling and Model Context Protocol (MCP) are transforming the financial advice profession.

They discuss why many advice firms are moving away from traditional all-in-one platforms in favour of best-of-breed technology, how MCP is creating seamless integration between advice tools, and why the combination of AI and deterministic modelling has the potential to dramatically improve advice quality, efficiency and client outcomes. Andrew and Ivan share practical examples of how advisers are using AI-connected modelling tools today, the impact this is having on paraplanning and adviser workflows, and what the advice technology stack of the future may look like.

LISTEN

SHOW NOTES

Topics Discussed

The evolution of financial advice technology
Why advisers are moving away from all-in-one platforms
Best-of-breed software versus legacy advice systems
The limitations of traditional modelling tools
What Model Context Protocol (MCP) is and why it matters
The difference between probabilistic AI and deterministic modelling
How AI can support financial modelling without compromising accuracy
Using AI to reduce advice bottlenecks and workflow friction
Real world retirement village modelling case study
The future role of advisers and paraplanners
How AI is changing the economics of advice delivery
Reducing the cost and time required to prepare advice
The three major bottlenecks in the advice process
Why financial advice firms need to rethink technology architecture
The future of integrated advice ecosystems
How firms can begin their transition to AI-enabled advice workflows

Episode Highlights

(Timestamps are approximate)

[00:00] – Introduction to Andrew Gardner, Ivon Gower and RetireMap
[01:20] – What traditional advice technology gets wrong
[04:10] – Why best-of-breed technology is gaining momentum
[07:20] – Transitioning away from legacy all-in-one platforms
[09:00] – Creating adviser-friendly modelling experiences
[12:50] – The retirement village case study and AI-assisted modelling
[16:00] – Deterministic versus probabilistic AI
[19:30] – The three advice choke points explained
[23:45] – When AI becomes transformational rather than superficial
[26:50] – How adviser and paraplanner roles may evolve
[29:40] – Taking back control of the advice process
[32:50] – Can legacy platforms keep up with AI?
[35:45] – Why modelling has been such a difficult problem to solve
[39:15] – What the advice technology stack looks like in five years
[43:20] – The impact of MCP on advice software integration
[45:20] – Final advice for practice owners considering AI adoption

Quotes

“The technology seemed ancient. It wasn’t keeping up with other industries, and it was practically impossible to use effectively with clients sitting in front of you.” – Andrew Gardner

“The simple test is this: is AI helping you write about your work, or is it helping you do the work?” – Ivon Gower

“AI shouldn’t be doing the calculations. It should be helping advisers leverage deterministic tools that produce repeatable, auditable outcomes.” – Ivon Gower

“The adviser gets control back. They no longer need to hand work off and wait days for answers.” – Andrew Gardner

“Once you can model multiple scenarios without a time penalty, the quality of advice improves dramatically.” – Ivon Gower

“If you’re not riding that AI horse, you’ll be chasing it.” – Andrew Gardner

“This isn’t hype. This is one of the first technology shifts I’ve seen that genuinely changes the economics of advice.” – Rob Pyne

Key Takeaways

The traditional all-in-one advice technology model is being challenged by specialist best-of-breed solutions.
MCP is creating a new level of connectivity between advice tools and AI platforms.
Deterministic modelling remains essential for financial advice because calculations must be accurate, auditable and repeatable.
AI becomes most valuable when it works alongside deterministic modelling rather than replacing it.
Advisers can test more scenarios, answer more client questions and produce better outcomes without significantly increasing time or cost.
The three major advice bottlenecks remain fact finding, modelling and SOA generation.
AI-enabled workflows can dramatically reduce delays between these stages.
Paraplanners are likely to evolve from software operators into strategy and quality specialists.
Advice businesses that embrace AI early may gain significant operational and commercial advantages.
Technology is making it increasingly viable to provide advice to younger clients and the next generation of wealth holders.
The future advice technology stack will likely consist of connected best-of-breed solutions rather than a single end-to-end platform.
Firms that embrace AI and MCP-enabled technology today will be better positioned to compete over the next decade.

Resources & Links

The Trusted Adviser: https://thetrustedadviser.com.au/
RetireMap: https://retiremap.com.au
Connect with Rob Pyne on LinkedIn
Connect with Andrew Gardner on LinkedIn
Connect with Ivon Gower on LinkedIn
Follow The Trusted Adviser Podcast

TRANSCRIPT

Rob (00:01.353)

Welcome Andrew Gardner and Ivon Gower to the Trusted Adviser Podcast.

ANDREW MURRAY GARDNER (00:04.785)

Great to be here, Rob, and thanks for inviting us.

Ivon (00:07.182)

Yeah, thank you, Rob. Good to be here.

Rob (00:08.959)

Thanks for joining me, guys. We’re having this conversation because we’ve been talking recently, especially you, Ivon, with some of our team members, and you and I, Andrew, about RetireMap more broadly. And the reason that’s come about is because I’ve been hearing about you from other people. So that’s kind of always a good sign, isn’t it? When people I know and trust that are actually in our industry are saying to me, Have you heard of RetireMap? And I’ve heard it from someone else just recently.

They say, well, we’re using it. And then people I’ve kind of built a really strong relationship with over the years have all been telling me that they’re looking at it or they have now started using RetireMap. So I thought, well, there’s a story here and I want to know what it is. So we’ve been chatting a little bit of late, and I think this is a great story to share for our listeners. We’re always looking for things to do to improve the way they go about their business. So we’ll kick off with a bit of background. Andrew, I’ll start with you. You’re the founder and MD of RetireMap. You spent years building a finance and financial planning practice before you founded RetireMap. What did that experience, your history, teach you about what software in this space was getting fundamentally wrong?

ANDREW MURRAY GARDNER (01:20.543)

Thanks, Rob. Looking back, it’s interesting. I started in the industry in 2000, then started my own business in 2002. And the first thing I noticed, particularly when I added financial planning to the business, was just the technology seemed to be ancient. It seemed not to be keeping up with other industries and other professions where they had really good technology, but I found the technology particularly for modelling, which is the space I work so much in, was just difficult to use. It was practically impossible to use in a client-facing situation. I always like to collaborate with clients and invite them to join me on that journey and go with them on the journey rather than just pointing into the future and say, I’ll meet you there.

I always thought that approach lacked a little bit of depth and it didn’t engage the client in the process, and by engaging them in the process, you can then have them take a greater sense of ownership. There were end-to-end technology platforms around, as there are now, and it was just hard to learn and to try and drive them, particularly when you’re in a client-facing situation, it was just hard to do because it was so big, so monolithic. I was talking to someone I know when I first started in the tech space about six years ago and and I spoke to a demonstrator of Xplan, I think it was at the time, and they said most advisers only know 10% of what Xplan is capable of doing. And I thought, geez, if it’s that hard to learn, how can you possibly do that in everyday life? How can you do that where you actually use it to bring a client with you? It took a long time to learn, a lot of expertise required to learn it, and of course

The by-product of that was that those who really were the experts back in those days and probably still are today are the paraplanners. And so the adviser was dependent upon the paraplanner to do it for them. Then you get back to the space of the adviser working with the client and then they change their mind. Or what about this? What about that? And then all that reworking having to be done again.

Rob (03:26.259)

I think a lot of people would hear that and say, Yep, that sounds about right. So, you speak about the end-to-end software and that the monolithic software, if you like, that tries to be all things. And because, ideally, you would think everything’s integrated into one platform, but that longstanding default of advice firms trying to run everything through a single platform is changing, isn’t it? the shift towards modular based tools that actually do a specific function really well is something I’ve noticed in talking to people

As people have started to say, Well actually, I want best-of-breed. I want things that actually kind of can play together nicely, but actually do one thing exceptionally well. What has changed in the market to make the all-in-one model start to feel more like a liability than a convenience?

ANDREW MURRAY GARDNER (04:11.817)

It’s interesting. I’ve been thinking about this a little bit over the last several months as we started to build out into the more seamless integration. I came to think that the train analogy was a pretty good analogy. You see those old country trains rambling down the tracks and they’ve got carriages and those carriages are joined together by a coupling down the bottom. But to get to the next carriage you’ve got to basically step outside, step across to the other platform and go inside. And that’s kind of what it’s like with some of the all-in-one platforms where you’ve got to move to that next part of the chain and then you’ve got to retrieve data from the from the previous platform and it’s just time consuming and you’ve got to have that expertise to know what buttons to press which order to press them in where to go to look for it and then bring it forward to do what you want to do in this in this modelling scenario. Whereas the best-of-breed approach it changes things a lot particularly with the introduction of AI and the platforms that are bringing that AI, whether it’s Paradino, whether it’s Claude, whether it’s Marloo or whether it’s Copilot, they create this new sense of seamless integration. And MCP, Model Context Protocol, has had a lot to do with that. In fact, it’s facilitated it. Whereas the traditional API approach, which I know is still used extensively, but that is mapping every single field and mapping it to the corresponding field in another platform and then transferring the data from one platform.

Platform to the other and bringing it in, and then if you miss a field or the fields changed, or if you haven’t maintained it, you might only get 50% to 60% connection. 50% to 60% of the data being brought across, and some are lower than that. On a good day, you might get 80 or 90 percent. But MCP allows you to bring across in as a complete bundle and bring it across in a way where it’s much more seamless and certainly complete, and in doing that, it makes it a whole lot better for the adviser to be able to bring that data across and know they’ve got the data. So, the analogy is that suddenly it’s an or it’s a very seamless integration, and each of those carriages are linked together with an open format where that data slips seamlessly through each of those carriages and brings it forward. You don’t need to go and retrieve data because it comes with you as you move through the process. And the other benefit, of course, is that it’s any component that’s not working for you, you can pull that component out and replace it with another.

ANDREW MURRAY GARDNER (06:41.377)

Best-of-breed component that works for you and your circumstances.

Rob (06:45.279)

Yeah, for sure. And that’s something that we certainly thought about back in the day when we moved away from the all-in-one solution that pretty much all other businesses tend to use as well. We did that back in 2017, looking for exactly what you’re describing there, Andrew, which is that best-of-breed things that will work together but actually are doing one thing exceptionally well. Ivon, let’s turn to you and to get this question around how the shift of best-of-breed happens. It doesn’t happen overnight. What does the transition period look like in a practice? That’s trying to move away from the all-in-one solution. What surprises firms most when they start to decouple from what platform they’ve relied on for years? How do they get it how do they go about that?

Ivon (07:25.749)

Well, I think there’s two things in there, Rob. The one is the technical side of it and the good news is it’s not a big bang migration.

You don’t need to move everything at once. The nature of the tech these days means that you can start to pick up a particular element of your advice process, one you feel isn’t really delivering you the efficiencies or the benefits that you’re after. And you can implement that and gradually migrate out of one of the bigger, all-encompassing systems into that best breed concept, right? So it’s phenomenal. We see people really expecting this to take a long time. But the reality is these days, because you’re working with such small specific elements of the process, you can genuinely start to receive benefits inside your advice workflow within just a couple of weeks as you start to adopt the new system.

Rob (08:27.401)

And this is something that Andrew’s alluded to there is being able to, I guess you can test drive it. As you say, you can you can move away from that component or that module of the all-in-one solution and start to test drive this function, which for you guys is about the modelling. It’s about replacing the cashflow modelling tool. And actually, as Andrew stated a bit earlier there, just being able to do that much more seamlessly and easily with clients. So being able to interact with the tool in front of a client if necessary, because the handoff if they do make a change to their strategy and you’ve got to go back to the paraplane team and that sort of delay in getting the job done makes this sort of the usability of something like RetireMap so appealing to advisers, I imagine.

Ivon (09:07.563)

Yeah, absolutely. This was something that amazed me when I first started working with the business. It’s the effort that’s gone into making something very complex, very intuitive. Even before we had the MCP connection, it was an easy system to drive very complex calculations and bring back a nice sort of simple interface that you could then share with clients and do your modelling on the fly.

Rob (09:32.682)

Yes. So, I’m gonna play devil’s advocate here a bit. I’m just gonna say that we’ve used Xplan for many years. We’ve got people in our business that are masters at WealthSolver or Xtools+. And there’s familiarity bias there. I’ll try that again. There is familiarity bias there and there is therefore probably a reluctance in some businesses for the experts to move away from what they’re familiar with advisers who’ve mastered a tool over the years don’t want to unlearn that and start something new, thinking this isn’t broken. They have gone to the trouble to become masters at Xtools+, which can do a lot. And I guess in that sense, they feel why is that there a need? And I know an adviser, in fact, not in our business, but he tells me he actually does update Xtools+ modelling in front of a client. And I think most advisers will look at it and think, how does he possibly do that? Because obviously he has put the time in to become an expert at it.

How do you get people past that, particularly when there is sort of that one person in the team that resists it and says, we’re used to this system, this works, why will we change? How do you get people past it and to test drive something that actually is a newer way of doing things but perhaps has a bit more usability for more than one master adviser?

ANDREW MURRAY GARDNER (10:46.463)

I think sometimes you’ve got to move leave the past in the past where a way of doing things isn’t always the best way and if it isn’t broken don’t bust it. Well there’s also a saying that if it isn’t broken maybe you need to break it to get to the next level.

Because in the end, time and efficiency is the really important part of advice. With AI coming in now, I was talking to one of our users, a high-level Melbourne user, from a high-level practice, and he was saying that he had a surgeon come in recently and he came in armed with Claude and a report that Claude had produced and also telling him what he should be paying in fees. And those fees were something like $15,000 under what. That advice practice would typically charge for something of such complexity. And so he got around that on that occasion, but it was certainly a warning bell for him, a red flag for him, that they need to become more efficient, more effective in the way they produce advice and the speed with which they can do that. And they’ve also got to take some of the cost out of it. PwC also had a write up in the financial review earlier this year and they’re talking about reducing one of their advice departments from 40 down to six, I think it was, and they’re producing the advice they’re providing to their clients faster, quicker, easier, and with more accuracy using AI as part of the solution. And I think that in this day and age where people are so well aware of what AI is doing, it’s on the news every day, it’s on TV every day. We’re all using it every day, and it makes things so much faster, so much quicker and so much easier.

Where things have been done. In the end, no money is made by an advice practice until the SOA is generated. And I think Ivon, you’ve got a great example there where one of the advisers came to you and says, I’ve got this problem with one of my clients going into a RetireMap village and they’ve got three or four different options plus other sub options and you solved that much more quickly than would otherwise have taken. Tell us about that one.

Ivon (12:50.615)

Well, yeah, this has got to be my favourite experience I think in over 20 years in the industry. A relatively complex scenario with an elderly client moving into a RetireMap village and she’d sold a home, had three different entry options into the village, each a different price point with different ongoing maintenance fees, and then this exit fee that applied across the first five years that she was there. The modelling itself was relatively awkward, but we sat there with Claude open on one screen, RetireMap open on the other, and were just entering prompts into Claude to see how we could model the first of those three scenarios. It was doing all the heavy lifting for us, but we could just look at the outcomes. And once were happy with that, the instruction was simple: just clone this. There’s two different variables, the entry fee and the ongoing fees and create two different alternative scenarios. So, it did that for us, a matter of minutes to then spin up the alternatives but the next piece was the analysis, and this is something that in years of doing paraplanning work would have taken me days to fulfil. We agreed between us that the unknown in this situation was the client’s longevity. How long would she stay in this RetireMap village? Would the exit fee apply, etc. So, we just verbalized this inside Claude. This is the unknown. Help us understand the implications of the time that she’s in there. We want to look at it across metrics like, cash flow, cumulative tax, cumulative age Pension, net worth, and bring us back the outcomes to help us really clarify this scenario for her. It produced in a matter of minutes nine pages of analysis, addressed everything we wanted. There was a matrix that it presented, three different entry options, four different timeframes, three, five, ten, and fifteen years and it gave us that clarity around that.

Ivon (15:00.227)

Decision point for her. But what really blew us away was it went further and it understood that this was not just a financial decision, this was a personal decision. This was where she’d be living. So it then gave us a list of questions to speak to the client about to help blend that optimal outcome of financial versus personal objectives. And what we got was a ready to review, ready to present, strategy paper that he could leverage.

ANDREW MURRAY GARDNER (15:31.232)

And going back to your original point, Rob, there, that one of the things that makes it easier to make that transition away from what one maybe already masters or their paraplanners masters, I think that there’s an opportunity here for us and we’ve picked up the ball with as a tech company to support for our users. So that they did come to Ivon for this assistance and he was able to help them do that. So, you’re not on your own. Suddenly you’ve got a tech company that provides assistance.

Rob (15:56.714)

Yes. It’s software but it’s actually supporting the use of the software where you actually are you actually got an expert on the other end of the line to say, this is what I’m trying to model out this scenario, can you help me work through it? And I think the thing is just to step back a second onto how you achieved that and you mentioned it, Andrew, the Model Context Protocol, the MCP functionality, the ability to use AI to do some of the lifting, but also, you’ve got RetireMap, which is a deterministic modelling software. It actually has a predictable output if you put the data in so you’re not getting a variable outcome using it, it’s mathematically sound time and again using your model, but you’ve used AI to connect to it and actually give it the context you needed it to and it helped to I guess amplify your ability to share the client’s scenarios to the RetireMap modelling software. Is that being that The MCP server connection between the two is what made that thing really work?

Ivon (16:56.577)

Yeah, I’ll take that one. I think that’s a hundred percent right. What we find is that MCP isn’t just about making it easier to do the work. It’s about supporting the analysis process and really helping with the modelling of variations on the advice or the what if scenarios that we’re very likely to come across when we sit down with a client and present the advice. That concept of meeting with a client, presenting advice, and then them asking the question that’s been sitting in the back of their mind, but they haven’t been able to really verbalize till they see the numbers. Whatever that is, what happens if I retire two years earlier? What happens if we want to leave an inheritance to the kids or help fund their first apartment purchase? Yes.

Rob (17:50.094)

Pick up on that Ivon. So essentially the AI component of that is the ability to effectively talk to the agent on the front side, though if you like, on the AI, whether it’s Claude or whatever AI tool you’re using, connect that conversation, if you like, to the RetireMap modelling software. It does the modelling work, but it’s actually the agent is interpreting the conversation the adviser’s having and punching that information through into RetireMap to do the deterministic modelling. Is that kind of how it functions?

Ivon (18:19.105)

That’s exactly right. Yes. So, there’s probabilistic and there’s deterministic, right? And probabilistic is the language model. That means feeding in a question, where the same question will give you different answers every time. It works quite well when you’re trying to construct content for an advice document or something because you can read it and you can review it and validate it, correct it, and be really comfortable that it’s done the work for you. When it comes to calculations, you want deterministic, you want to get the same result every time. Same inputs, same outputs. And that’s really important because if the client then asks a question, how did you get to this number? Or a regulator comes in and says, give us the substance behind your calculations and the basis.

ANDREW MURRAY GARDNER (18:56.321)

Okay. Ivon

For your advice, you’ve got that fixed, auditable, transparent list of numbers that you can fall back on. So blending the two becomes really powerful. And to sort of clarify, the MCP capability in RetireMap drives the work that you would do to input the modelling. It doesn’t run the numbers itself.

Rob (19:30.975)

Yeah, for sure. And Andrew, if I can turn to you there, just when we’re chatting about I guess the way technology is helping to change the way advice is delivered and the workflow is kind of being altered completely. You talked about three choke points in an advice practice. You said the fact-find, the modelling, and the SOA. You’ve clearly gone headfirst into the modelling space to address that key choke point which everyone can relate to based on what you both have just said. Which of those three has historically been the most neglected?

I guess from a technology investment perspective and why? Is it has it been the modelling for you? Is that the way you thought we’re gonna go into modelling ’cause we don’t think this is actually being properly handled, in a way that’s useful for the client, useful for the adviser and easy to model multiple scenarios at short notice?

ANDREW MURRAY GARDNER (20:20.181)

Yeah, that’s right. The modelling was the part that I found difficult in the practice. That if you don’t have something that you can use quickly, easily and efficiently and have the client engaged and see the outputs on it on the spot, then they can’t become really connected to what’s happening. But by extending it on either side, so the likes of Paradino and Claude and Marloo and the other AI agents and Copilot for that matter as well.

Those AI agents have allowed for the other two to open up as well so that you’d the data flows in so easily. Ivon’s done some lab testing and it’s real world testing as well, that within two or three minutes he’s able to issue a command to take to lift the data out of the AI agent and place it into RetireMap, but you’ve got a base plan then. Then you can run these scenarios running simple scripts, simple prompts. And you can run them on top of each other. And within that, you can then go to the plan comparison reports in RetireMap to then compare them against each other and see which of those is most closely aligned to the goals and objectives of the client and be able to present that as a graph so that the client can see it, the adviser can see it, and both of them have an end result that at the end the yeah the adviser has definitive instructions.

The client has run through all the different scenarios that they want, so they’ve got a definitive outcome and an expectation, and then you’re able to push that out into the into the AI agent to produce the SOA. So it’s extracted, it’s pulled out, but that gives you the end result of all three of those coming together, but the central part of it is being able to do that modelling quickly and easily. The data goes in without manual data entry. It’s just a prompt. And it’s then the end result is extracted out into the AI agent for SOA production at the press of a button. So all those things come together. But in the end you need to be able to do the modelling, and the mod modelling is central to getting the outcomes and and the reporting so that the client can see where they’re going. You don’t have any of those afterthoughts of what about this or what about that. And all this brings forward

ANDREW MURRAY GARDNER (22:34.165)

The end result is the SOA can be produced more quickly, provide it to the client, and the cash flow comes forward and everyone’s happy because someone’s told me just the other day actually that older Australians are the ones who are demanding immediacy about the outcomes because they don’t want to wait. And it’s a bit like Amazon got flown ahead and they deliver today or next day. If you can produce your SOA that much that much more quickly, then you’ve got an edge on the market and

This combination is giving you that edge where you can do the job more effectively, more efficiently and save a whole lot of cost in a range of areas.

Rob (23:09.437)

Okay. So Ivon, you’re the guy people are talking to when they’ve got a models modelling scenario they want to sort of test drive and they’ve got this one recently you talked about came to you and said I’ve got a complex case, aged care scenarios, multiple different things to model, and you therefore have got the ability to take that scenario, obviously you’ve got expertise with your tool as well as the AI at the connector that actually can make it work. Where does kind of AI move from being superficial, note taking, executive summaries, producing

SOA outputs that maybe aren’t necessarily accurate because people haven’t used a deterministic model to actually get the right consistency in their outputs. Where does surface-level AI end and genuinely transformative AI capability begin for you from what you’ve experienced with your clients and what you build?

Ivon (23:59.818)

I would say the simple test is the AI just helping you write about your work or is it helping you to do the work? So, you’ve got your note takers, you’ve got your tools that generate executive summaries for you and I think those are genuinely useful. They cut out time, they help improve the quality of content that goes to a client, but they kind of operate on the outskirts. Of advice. They then transformative is when the AI actually participates in producing the advice. You’re building a fact find, you’re running strategy comparisons through a deterministic engine. You’re helping interrogate why one scenario is more effective in meeting the client’s objectives than another. And it’s not just making you faster, it’s making the advice better. It’s improving the quality of the advice because it’s really supporting the end outcome of client understanding and

Rob (25:07.165)

think it’s also it’s better because people are prepared to model more scenarios because the time isn’t sort of a constraint anymore. They’re not thinking about I just can’t model another scenario on this one. There’s just too much time involved and we think we’ve got a close enough answer. Whereas the speed with which you can model scenarios using the AI interface, does that you think result in a better output in terms of the end result for the client, not just in terms of speed but in terms of quality of advice being delivered?

Ivon (25:34.754)

Yeah, absolutely. I think that once you have that luxury of being able to model as many different variations as you need to, you cover off all those questions. So in many cases the adviser understands the client strategy, the best strategy for a client, almost immediately. But proving that and then challenging that now becomes something that they can do for every client. Without a time cost associated to it. So I think yeah.

Rob (26:07.583)

Sorry to cut you off there. Do you feel that with what you’re now doing and seeing and helping people achieve with the software that you’re using, what does it say about where the profession’s heading for you? Roll forward maybe a year or two from now, what do you think the profession looks like with the ability to use AI connected to deterministic tools? And there’s a very distinct distinction you made there. I think it’s really useful to revisit.

Probabilistic, I think it was you, Andrew, that said this probabilistic versus deterministic. Probabilistic is that sort of AI interface. You can run multiple scenarios. Deterministic is the output making sure it’s consistent, mathematically sound every time. But let’s roll forward a year or two. What do you think changes in the way advice practices are running compared today?

Ivon (26:54.933)

I’m taking that one. I’m very happy to. I think that the structure of the advice practice might start to change. Right. What I see happening inside businesses that are using AI really effectively is that everyone sorts of steps up a level. You used to have your paraplanners being experts in technology who could get the outcomes that advisers wanted in a projection tool. Now that’s easy, so the paraplanners can be strategy experts, can be people who are really sort of testing out and fine tuning the strategy to get the best result. You used to have Yes. Rob (27:35.126)

Do you see paraplanners sorry, I’ve jumped in again there. Do you see paraplanners stepping in and actually taking on a co kind of a quality assurance role, helping to make sure the tools are working as expected? They’re constantly looking to iterate on what’s being used to improve upon. So, they’re becoming more process oriented and outcomes oriented rather than just generating advice documents. They become more sort of custodians of the way in which work is done.

Ivon (28:02.125)

Potentially. I try to frame everything in the context of client outcomes. So I think that to the extent that we can use those paraplanning roles to understand the workflow and produce a quicker, better quality outcome, then absolutely.

I think though that the paraplanners can also sit across the quality of the advice itself and the client experience itself. They can be the ones who look at those various iterations and say, here are some opportunities that you might not have thought of. We’re not just modelling to meet your objectives, we’re modelling to exceed your objectives. And here are some scenarios that we really want to engage you with to consider that were potentially beyond your expectations.

Rob (28:47.891)

Okay, so let’s go to this point about roles changing and perhaps some roles changing dramatically, such as the paraplanning role. And I think what you guys are describing very much points to that fact. Some businesses are saying, well, you don’t need to send it to the expert paraplanner to write the document anymore because you can actually model those scenarios. You’ve got the strategy ideas in your head, you can simply verbalize those strategy ideas, you can run multiple scenarios, you can actually get the output you want without relying on the expert paraplanners who knows X Tools plus like the back of their hand that can actually model that scenario. So what would you say to an adviser then that’s actually worried about AI making them redundant? Others can see others can’t see it and say, well I don’t think it’ll change our day at all, the way I work at all. What’s the reality or where’s the reality for the adviser in say, running a firm, 10 or 20 people over the next three years, what do you think the adviser’s role looks like with the changes that are coming down the pipe?

ANDREW MURRAY GARDNER (29:42.41)

I think the biggest change is gonna be for the adviser is that the adviser’s gonna take control back. They’re gonna risk control back from the external sources that have been doing this work for them. So they’ve been sitting there talking to the client and discussing different ideas, taking notes of those, and then they’re writing the brief. We spoke to just we spoke to a young adviser recently who’d be come out of Macquarie and BTs. They had very high-level clients, very complex structures. He explained that he spent a day, I could hardly believe it myself, a day writing the brief for the paraplanners to go ahead and model what he wanted to model. He showed Ivon the flow chart of this incredibly complex structure. And he asked Ivon how long would it take you to run this through? And Ivon said two hours with confidence. And he said, geez, this takes me all day just to write the brief for the for the paraplanner. Then I’ve got the queries for it. So I see the role of the adviser changing from regain control. So retaining that control, getting it back into their hands and being able to

The entire process to cut their time down of all this backroom work of being able to do it actively in front of a client and give them the ability to service many more clients and do it more cost-effectively because they haven’t got this outsourcing. Even this morning I spoke to a practice with just two advisers, a very small practice. They’re spending $74,000 a year on external paraplanners, and they see being able to bring this back into the office as being a way of retaining control, getting that control, resting control back is what I’m trying to say. Wresting control back. And being able to contain their cost, deliver advice more effectively, more fluidly, and more effectively in the sense that they can cover off any of those things that the client comes up with and they don’t have to rework anything. So cost control, getting control back and then being able to spend more time

ANDREW MURRAY GARDNER (31:50.037)

Doing what they enjoy doing and that is sitting in front of a client and discussing the options and running those scenarios and giving the outcomes.

Rob (31:57.822)

Yes. I think everyone listening would totally relate to what you’re describing, Andrew. I think it’s the promise of what it will do for businesses is quite profound. It’s nothing that I’ve ever seen. We I’ve been going since about the mid-nineties in this profession in one form or another. And it’s the one area I always thought technology was kind of overpromised, under delivered. This is one I think where the hype is not just hot air. There’s actually some real practical application that people are using right now to materially change the economics of their business. Let’s go to the sort of the incumbent. Xplan has somewhere around 65% to 70% of the Australian market. There’s a version of the innovator’s dilemma playing out there for those that haven’t read the book. Clayton Christensen wrote a great book called The Innovator’s Dilemma, where rebuilding legacy architecture at that scale that Xplan has built is almost impossibly expensive and risky for them. Do you think incumbent platforms like Xplan can actually respond to what’s happening?

Or is the trajectory already locked in, do you think, for them?

ANDREW MURRAY GARDNER (32:59.187)

It’s interesting. I was speaking to an IT consultant who worked with large companies. He actually worked in Telstra, and I asked him about these large platforms and where they’re at with what they’ve got to do to get into the AI space. And he said to rebuild a platform that is very large and it’s got it’s got legacy architecture that’s years old.

It’s going to be a tough call because there’s a lot of money that needs to be invested into it and it’s going to take a lot of time to do that and a lot of planning to do that. But the big part is not even that, the big part is migrating from the old tech to the new tech to connect seamlessly with the same architecture into AI agents. We’re very fortunate that we made the fortuitous decision at the end of 2024 and through 2025 to rebuild our architecture from the ground up, and we engaged a team of software experts, which fortunately we did because they built it on the same architecture that the AI agents today are building theirs on, which allows for us to have the same plumbing, the same piping, the same architecture as they have, which makes a seamless integration, it just moves so fluidly. But if you take architecture that might be eight years old and in some cases maybe older, and for that piping, that plumbing to fit into modern architecture, I think it’s a real challenge for them. I don’t know, but I think. It might be a real challenge for them to do it and do it in a seamless manner.

Ivon (34:29.133)

I think just to add to that, the practical implementation of a change like that. We talked at the start around this all-in-one technology. The question would be if you are managing the end-to-end advice workflow, where do you start? What’s your first introduction, your first foray into the AI supported workflow? I would venture to say it’s not going to be the cashflow modelling piece because that’s a pretty challenging one too. It’s a it’s a hard nut to crack. So when you’ve when you’re managing the end-to-end piece, you’ve just got to bite off little pieces at a time and it’s gonna take a long time to get there. The beauty of the best-of-breed functionality is you’ve got specialists just focusing on delivering that solution in their space.

ANDREW MURRAY GARDNER (35:21.285)

And to that point, to that point Ivon I was speaking to an adviser recently who said that one of their platforms, end-to-end platforms, said they’d introduced AI and the AI was actually writing an executive summary of the SOA, hardly AI. Another one introduced AI by announcement and effectively what it did was AI notes, which has been around for a while. So you start to wonder what introducing AI really is.

Rob (35:21.289)

Yeah, it is that

Rob (35:46.9)

Yeah, and just to go to that point you made there, Ivon, about it being a hard nut to crack, the modelling piece, it’s it’s an area that when people have been looking to, I guess, look for alternatives to the large incumbent that’s in the market, the modelling has always been the one people feel like I you just can’t get the robust depth of modelling that you get from Xtools+. You just can’t find it. Other software has been out there to build something that is better and more presentable, I suppose, in the sense of client-engaging, in terms of client engagement, but perhaps not the degree of complexity that was built into the Xtools+ software platform. You’ve spent six-and-a-half years building RetireMap. And you as you said there, Andrew, a moment ago, you’ve rebuilt it to actually build it on the platform that now enables AI to interact with it, to actually connect it, to make it much more, I guess, leverage the advisers’ time in ways that advisers can’t possibly imagine until they start to use it. As you say, it changes the entire workflow. What does it say, I guess, the fact that you spent this time on the modelling piece? What does it say about the competitive landscape that this area remained largely unsolved for so many platforms that it couldn’t get the same level of capability that Xtools+ had?

ANDREW MURRAY GARDNER (37:04.533)

Yeah, that’s that’s true. And it seems like Xtools+ has owned the space for a long time because they did have the depth. We made the fortunate decision back in the early days to start with the more complex part of modelling so we started with entities and we built downwards into Centrelink from there and that gave us the right architecture at the top end to be able to push down into those lower ends. It’s harder to do the complex when you’ve started with the with the less complex but doing the entities was something were really keen to really build out and to cover every option even Division 7A loans and the ability to

Rob (37:43.059)

What made you start there, Andrew? Why did you start there when most people wouldn’t have started there? They have started with the simple ones and then build it out from there. Why did you start at that complex end?

ANDREW MURRAY GARDNER (37:51.264)

We wanted to have a program, a cashflow modelling program, that had sophistication, depth and substance to it. And we felt the best way to do that was to work in the in the space where there was least competition in that space, but also something that we enjoyed doing that sort of substance. I’ve already always enjoyed working in structures. I’ve been doing that for nearly 30 years. And it’s something that I like working in that space and I have been in that space for a long time. So we felt that was a good place to start working and to build a name for a time map as something of substance and to what to build down into the other space, and it took us years to do that. And the other part that took a long time too was just to get the right mix of granular detail and then summarize granular detail and then the graphics in the visuals so that the adviser can follow where that chain of calculation starts and follow it all the way through to give a summary so the client can understand where it is, and then to give the visual. So that the client can see the trends and the movements and to be able to move that from present values to future values putting inflation in and out all those sorts of things are all new challenges to do and there are so many different parts of the equation and modelling to handle and that’s why it has taken six-and-a-half years to get there.

Rob (39:12.629)

Yes. Ivon, tell me, as as a technician who’s helping people use RetireMap, let’s you might have answered this question already, but I think it’s worth just spending a second further on it. Let’s go out five years. What does a well-run advice practice’s technology stack actually look like? Is it this best-of-breed tools all connected via MCP server, do you think? Or what’s your what’s your view?

Ivon (39:39.616)

In five years it could be the next evolution of MCP, but the short answer is yes. I think the ability what’s held everything back in the past has been the connectivity, right? Everyone looks at the single purpose tools and says these operate incredibly well, but you’re asking me to duplicate my data entry. You’re asking me to transfer data from one system to another and then it’s not coming back. The MCP capability will just enable this flow of data to move as quickly as the ideas do for a practice. So you have a connected CRM, a connected digital fact find to collect information from clients, the modelling engine, the SOA generation tool, then add revenue management, your invoicing, your client communications. All of these tools can sit together and speak to each other. So the ability to simply sort of say to your CRM, find me all clients who will retire in the next five years but have a potential funding gap. Now

Rob (40:48.415)

MC yeah, MCP’s been the real unlock there, hasn’t it? That’s because without MCP you had the situation where you had best-of-breed but you had to build API connection and as Andrew mentioned earlier you might get not all that data cross or if you get it then you change the model or they upgrade their software and then the it breaks the API or you get half the data you used to get. So MCP has been this is the kind of key point here, isn’t it? It’s been the real unlock. It means that if software is built.

Ivon (40:51.624)

Absolutely. Yes.

Rob (41:15.645)

Genuinely with the ability to connect to other parts of the ecosystem through MCP, it means that people don’t need to think I have to have an all-in-one solution because that’s where my data has to go to get the plan out the other end. I’ve got to model it there because it’s where the data is pushing through into my SOA production engine. Now you’re saying you don’t need to. You can do that over here and you connect it to your plan production software over there, and it actually works seamlessly. You don’t need to necessarily have this all complete solution anymore. It’s very much modular. Connected via MCP.

Ivon (41:47.308)

Spot on. Yes. And MCP brings the flexibility. Client scenarios change, practice needs change. MCP is not that structured integration between two systems. You don’t need to think through all of the different ways that they’ll connect and then continue building as new needs open up. MCP gives the agent the instruction manual. To leverage all of the functionality that sits in the system. And the beauty is that as different tech businesses continue to build and enhance their functionality, the MCP just moves with it. So you don’t need to continue to coordinate various integration partners in order to get the APIs continuing to be built.

Rob (42:31.871)

Yes.

ANDREW MURRAY GARDNER (42:31.935)

And to that point too, Ivon, as you as you do on a day to day basis, not only are you using the MCP for your SOA generation and your note-taking and that integration there, but you’re also bringing Claude into the equation or other AI agents to be able to add to the equation as well and to add different dimensions, as you did with that RetireMap house scenario where you brought Claude into the equation, that all the work was still being done deterministic in RetireMap, but you were able to use those different agents because MCP is that general widespread broad connector across AI agents. So if it’s got an AI agent, if it’s got a chat capability, it can connect to RetireMap MCP and other MCPs to give the full range of needs, regardless of what they are. And I think that’s that’s a major change. And this has really only started in the last quarter or so, Rob. It’s not like this has been going a long time. We’re talking only a few months that this has been able to happen and it’s changed things dramatically.

Rob (43:27.669)

Yes.

Rob (43:32.583)

It’s changed the game for sure.

Ivon (43:33.024)

My un my instruction to every person that uses the MCP is once you’ve given it an instruction, add the question. Ask me if you’ve got any questions or clarify if you need anything further. And invariably the agent is able to come back and say, I need more information on this or that or something else. And this is the scenario where you’ve given information to a paraplanners and ask them to produce an SOA. And they come back three or four days later and say, I’m working on this now. Can you give me some extra information? This is immediate and it helps to round out the instructions you provide; it helps to fill in, so absolutely.

Rob (44:17.685)

Well, it’s so much more efficient too, because you’re actually then doing it on the spot at the time, not handing it off and then getting asked questions three days after you’ve moved on. So you’re actually doing it all at the same time. So, it makes a huge amount of sense. And I think for those that haven’t yet, I guess, used the tools the way we’re describing them here, there is actually no turning back now. Whether it’s whether it’s a version of MCP that’s evolved in five years’ time, the methodology of connecting tools this by this method or a version of that

Ivon (44:25.355)

Yes.

Rob (44:46.217)

That it changes to over time; it just creates this seamless integration between best-of-breed tools as we’ve described. So, I’m super excited about the prospect of where things are going for all of our businesses because there’s more people that want to get advice, more people we should be serving out there. And it’s just been hard to deliver it in an efficient way from a time and cost perspective. But for you, Andrew, one final question if I may. For a practice principal sitting there right now, still fully embedded in Xplan. And not quite sure where to start, what’s the one thing you’d tell them to do first?

ANDREW MURRAY GARDNER (45:21.547)

Financial advice has been starved of advanced technology for a long time. Suddenly it’s arrived. The thing they need to do is to embrace it because it’s going to change the way their practice operates, it’s going to change the way they allocate their resources. It’s going to be it’s going to remove those three choke points. It’s going to allow advice to free up and move more seamlessly and more smoothly through the entire process of advice. And you as the owner of a of a large practice servicing very high net worth and ultra high net worth clients, I’m sure you’ll agree with the others we work with in this space, that they’ve got they’ve got children. Children with smaller super balances, smaller FUM, and suddenly they can afford to service them because if they can knock out an SOA and do it effectively and do it in a single meeting with the client and do it for a few thousand dollars and they’ve only spent an hour, maybe an hour and a half or two, they’re still getting well paid for it. They’re no longer doing it as a job just to ring fence the kids of their high value clients, but they’re doing it because it’s actually profitable work, embrace the technology and to move with the times.

I heard a professor of AI on Talkback Radio here in Melbourne just a couple of weeks ago and he says if you’re not riding that AI horse, you’ll be chasing it and you may never catch it. Get onto it now and embrace it because it’s going to make a world of difference in the efficiency and effectiveness of practice and advisers and make everything just work so much better and smoother.

Rob (46:55.925)

Yeah, I couldn’t agree more. It’s exciting to think that we’ll be able to do more of the work we’ve been trying to do for as many people as possible but do it in a way that just gives us that opportunity. And as you say, the intergenerational wealth transfer has been talked about for a in a lot of in the media in the l a lot in the last couple of years, and I think the ability to actually serve a younger audience perhaps that actually doesn’t have the wealth yet but can it can be done cost effectively to them as well. It’s certainly very promising. I really appreciate you both spending a bit of time with me today. We obviously got to talking because as I said, some people I know and trust have been in this profession a long time like I have been saying they’ve already used your tool, already are using it and have never looked back. And that’s a good sign. They’ve not got regretted it and thought, hang on, we’ve now lost this functionality, but they actually haven’t looked back, and they’ve said we’ve

We’ve really embraced it and it’s made our job a lot easier and now they’re actually adapting to the to the AI capability that goes with it. So Ivon Gower, Andrew Gardner, thank you both for joining me today on the Trusted adviser podcast.

ANDREW MURRAY GARDNER (48:01.355)

Yeah, pleasure. Thanks for having us, Rob.

Ivon (48:01.517)

Thanks so much.