EPISODE 51- The Data Foundation Behind Modern Advice: Stephen Handley of Fin365

By /

In this episode of the Trusted Adviser Podcast, Rob Pyne sits down with Stephen Handley, Founder of Fin365, to explore the evolution of advice technology, the importance of data ownership and why financial advice businesses need to get their foundations right before embracing the latest wave of AI.

Stephen has spent more than a decade watching the advice industry attempt to modernise. What began as an internal technology project inside his own financial planning business has evolved into Fin365, now used by 174 advice firms and around 2,000 users.

Stephen shares the story behind that journey, including why he chose Microsoft Dynamics over traditional advice software, why Fin365 deliberately took the harder path of building direct integrations with wealth platforms, and how the business has evolved from a software company into a broader technology and services provider following its acquisition of Quirk.

The conversation also explores one of the most important questions facing advice businesses today: what should come before AI?

Stephen argues that firms should think about technology like building a house. Before adding AI tools, client portals or automated workflows, they need a strong foundation – clean, structured and connected data, clear permissions and genuine ownership of their information.

He also demystifies APIs and Model Context Protocol (MCP), explains how Fin365 is experimenting with AI agents, and discusses why deterministic systems still have an important role to play when accuracy and consistency matter.

LISTEN

SHOW NOTES

Topics Discussed

 

  • Stephen’s journey from financial planning into technology
  • Building an internal technology solution for his own advice business
  • Why Stephen chose Microsoft Dynamics as the foundation for Fin365
  • The shift from all-in-one advice platforms towards open technology ecosystems
  • How greater independence has changed technology decisions for advice firms
  • Spinning Fin365 out of the advice business in 2016
  • Growing Fin365 from a bootstrapped startup into a technology business serving 174 firms
  • The challenges of scaling software and onboarding new advice businesses
  • Why Fin365 chose to build direct data feeds rather than rely on aggregators
  • The importance of data ownership for financial advice businesses
  • What advisers should ask software providers about data ownership
  • The difference between owning your data and simply having access to it
  • Why Fin365 deliberately avoids trying to become an all-in-one advice platform
  • The acquisition of Quirk and the growth of Fin365’s technology services offering
  • Helping advice firms customise workflows, reporting and technology
  • Why technology services are focused specifically on the Australian financial advice sector
  • The changing advice production technology landscape
  • Why SOA automation is only one part of the efficiency equation
  • Deterministic modelling versus probabilistic AI
  • APIs versus Model Context Protocol (MCP)
  • Building an investment analyst AI agent
  • Why Fin365 is taking a hybrid approach to AI
  • The importance of clean and structured data before implementing AI
  • Why advice businesses should start with the foundation rather than the latest technology
  • The future of AI and technology in financial advice

Episode Highlights

(Timestamps are approximate)

  • [00:00] – Introducing Stephen Handley and the story behind Fin365
  • [02:20] – From financial planning to building a technology business
  • [04:30] – Why Stephen chose Microsoft Dynamics
  • [07:20] – The evolution from all-in-one platforms to technology ecosystems
  • [11:00] – Spinning Fin365 out of the advice business
  • [13:30] – Growing to 174 firms and thousands of users
  • [17:20] – Scaling the business and acquiring Quirk
  • [24:30] – Why Fin365 chose direct platform integrations
  • [27:30] – The importance of owning your data
  • [36:00] – What advice firms should ask their software providers
  • [40:00] – Making sense of the crowded advice technology market
  • [44:00] – Deterministic modelling and the role of AI
  • [50:00] – Matt McGuinty and the growth of Fin365’s technology services
  • [56:30] – APIs, integrations and why they matter
  • [60:00] – Demystifying Model Context Protocol (MCP)
  • [63:30] – Building Fin365’s investment analyst AI agent
  • [68:00] – Why the data foundation needs to come first
  • [72:00] – Stephen’s advice for firms navigating the technology landscape

Quotes

  • “AI is the current shiny light that everyone’s paying attention to, but it’s really just a layer on top of the underlying stack.” – Stephen Handley
  • “For me, owning the data and making sure the data was as rich as possible was a core selling point.” – Stephen Handley
  • “If I leave the dealer group, I don’t have to ask permission for my data.” – Stephen Handley
  • “The full automation of a statement of advice to me was never the be-all and end-all of what I was trying to build with my technology.” – Stephen Handley
  • “You start from the ground up. You don’t go and start from the cool AI tool.” – Stephen Handley
  • “If you do not have the data foundation centralised, clean, complete, well organised, all of the relationships between the different data points are well structured, then you don’t have the foundation.” – Stephen Handley
  • “You’ve got to start with the foundation and then work your way up.” – Stephen Handley

KEY TAKEAWAYS

  • Technology starts with the data foundation. Before adopting AI or new automation tools, advice firms need clean, structured and connected data.
  • Data ownership matters. Stephen argues that firms should understand whether they genuinely own and control their data, rather than simply having access to it through a third-party system.
  • Open ecosystems can outperform all-in-one systems. Rather than trying to build every capability into one platform, Fin365 has focused on connecting best-of-breed technologies through an underlying data foundation.
  • Direct integrations can create long-term advantages. Fin365 chose to connect directly with wealth platforms rather than relying on aggregators, despite the additional complexity and time involved.
  • AI doesn’t replace the need for deterministic systems. When producing investment or advice-related outputs, consistency and accuracy remain critical. Fin365 is therefore exploring hybrid approaches where AI sits on top of structured, deterministic processes.
  • SOA automation isn’t the only efficiency opportunity. Advice businesses should look across the entire client journey and back-office workflow rather than focusing exclusively on advice production.
  • Technology services can be just as important as software. The right technology still needs to be implemented, configured and integrated effectively to deliver its full value.
  • Focus matters. Fin365 has deliberately concentrated on the Australian financial planning and financial services market rather than trying to become a general-purpose technology provider.
  • MCP is another layer rather than a replacement for APIs. Stephen explains how MCP allows AI agents to interact with existing systems without needing to understand the underlying API directly.
  • Don’t start with the shiny new tool. The biggest lesson from Stephen’s experience is to build the foundation first, then layer technology and AI on top of it.

RESOURCES & LINKS

TRANSCRIPT

Rob (00:01.676)

Welcome Stephen Handley to the Trusted Advisor Podcast.

Stephen (00:04.774)

Thanks, Rob. Appreciate you having me on.

Rob (00:07.992)

Thank you for staying up late in Texas. You’re at seven PM Texas time and I had heard from someone that you lived in the US and then I saw the Texas plates in your background and the Captain America Shield and thought, yep, it’s true. So, it sounds like you’ve got a good rhythm there and you don’t mind taking meetings. It’s this late in the evening for you, so it’s nine AM Perth time. but appreciate you staying up and catching up with me on this chat, Stephen.

Stephen (00:37.102)

No problem. It is yeah, for a while it was just a rumor. No one knew whether it was true or not, but it is true. I do live in Texas, so I do have a Captain America Shield behind me, and staying up and working afternoon shift is par for the course. So, this is my regular daytime. I used it once for as a prop for my wife’s birthday.

Rob (00:52.962)

Have you ever needed the Captain America Shield?

Stephen (01:04.164)

I won’t go into all the details, but it was I was basically making a a bit of a video it was during COVID and we were putting together one of those video montages ’cause no one could be at the party. And so, I used the Captain America shield as a prop, but that’s about it. Other than that, it’s just decoration.

Rob (01:15.757)

Yeah, right.

Rob (01:21.336)

I reckon it’d be good for a party in person, so maybe a future party, dressing up as a Captain America. Let’s get into the story behind Fin365 and your story. It’s a fascinating one because and I reached out to you or I had someone from my team reach out to you because every person it seemed I was talking to was saying they were already on Fin365 or planning to go to Fin365 was all queued up. So, I thought I’ve got to talk to Stephen to find out just what scale the business has grown to and what the backstory was. So, ten years in, you know, you’ve been I think when we spoke about your background, you said you watched the industry try to modernize more than once. You know, you have been you’ve come from the industry, you know, you were actually I’ll just go let me start it again, Stephen, because I was actually thinking I was gonna ask a different question there. So, I stuffed it right from the start.

Stephen (02:14.869)

Yeah, right. No worries. No worries.

Rob (02:18.732)

So, take us back to the start. You had a life in technology, you went into your father’s financial planning business as part of a succession plan that ended with the business being sold, and that pushed you into starting your own practice serving mums and dads. What did you find out when you went looking for software in your own practice?

Stephen (02:39.547)

So for me, having had the the background in software and tech, I was always just curious about optimizing any process, any workflow, any in terms of efficiency. So, there’s always just been a curiosity. And when I set up my own financial planning business, and as you said, we were serving mums and dads, and so our average revenue was probably half the industry average at the time. I wasn’t serving the high net worth or the mass affluent. And I am not naturally a micromanager. I didn’t want the business to be dependent upon me. I knew that if you know if I was going to grow this business, there would be other advisors who would come into the business and did come into the business. So, for me it was around operational affix efficiency, it was around automation.

I knew that meant integration of systems. And when I looked at the proprietary systems that were available off the shelf at the time, while they may have been they may have been good for advice production and the good old statement of advice that we needed to produce, I outsourced paraplanning, so I really didn’t care too much about how about my software being able to do the advice production piece. I was looking more at that back office automation and that led me down a pathway of exploring CRMs and eventually after a couple of different alternatives that I looked at, settled on Microsoft Dynamics as the CRM that I would use for my business. And that that was sort of the first five years of the journey was just using dynamics in my business and customising it for my own personal needs in the business.

Rob (04:33.634)

Yeah, and when you made that decision to go with Dynamics, you said that it was a bit clunky, but you were making a bet at the time, rather than doing a product comparison looking at the best available CRM, you’re really making a bet on the future of where you thought things were going to be going. What were you betting on at the time? And has it paid off the way you thought it would?

Stephen (04:55.289)

Yeah, so the answer to the second part is yes, it’s paid off in spades. I did look at others. I spent twelve months customizing sugar in the beginning. It was an open-source CRM. Didn’t cost me anything. So that was that was an immediate benefit. But it got to the point, and this was really the answer to your first question, it got to the point where I wanted to integrate sugar with Outlook and SharePoint and other business tools that were primarily Microsoft based and it was difficult and I just ran up against brick walls. I did look at other systems like Salesforce is an obvious one and there’s a number of other Salesforce options out there. For me, Salesforce was overkill. It was very costly to both license and develop. It’s got proprietary coding technologies that were required. It was much easier for me to find dynamics developers even back then and so the combination of balancing the costs, but most importantly, the integrations with the existing Microsoft tools. And I had worked at Microsoft. I’d had interactions with the business applications group over there. I saw the transformation that they did with when they way they took Microsoft Office to the cloud. And it was a bit of a lucky bet that if they continued in the direction that I had witnessed when I worked there and they got the integration of the CRM and all the other business applications seamless and working sort of best of breed, then that was going to deliver significant back-office efficiencies for me. And fast forward, we’ve got Power BI, you’ve got Microsoft Teams, you’ve got Copilot, you’ve got Power Automate, you’ve got Microsoft bookings. I mean, there’s just so many tools all in that one suite, all from Microsoft and for the most part, the integrations at this stage are seamless. So, it really has paid off.

Rob (06:52.024)

Yeah. I want to get into that a bit more detail. Before I do, this industry’s tried to modernize more than once. What’s foundationally different about the position advice business is in today compared to ten years ago when it comes to technology decisions they’re trying to make because it’s there there’s no more topical discussion than AI and technology. What do you think is different? It’s it AI or is there tell me about your perspective on the transition you’ve seen over the last decade.

Stephen (07:23.845)

Yeah, so I mean AI is the current shiny light, you know, the sparkly bauble that everyone’s paying attention to but it’s really just a layer on top of the underlying stack. And I think the real transformations come about for a few reasons. The first is the concept of an ecosystem didn’t exist ten, fifteen years ago. The proprietary systems that were available, the Xplans, the Advisor Logics, the Midwinters. They’re all an all-in-one solution, primarily mostly closed from the broader suite of office technologies. they did modeling, they had basic CRM functionality, they had research tools, they produced documents. The no one had really tried to take a different approach, which is focus, which is what I ended up doing, which is focus on the data foundation. And rather than try and build every application and every tool yourself, look to integrate with other tools that were out there. So that that was a new concept. There are others now doing that, but at the time no one else had. The other significant shift is not so much a tech shift, but just an industry shift in that ten, fifteen years ago, the majority of advisors were still under dealer groups that were dictating, largely dictating the software that could be used. When I first set up my financial planning business, we were under a dealer group and we had to use Xplan and as most advisors do, I used it to the bare minimum. I had my own CRM on the side because it met my needs better than Xplan did, but then eventually I got my own license and at that point that freed me up to make tech all the technology s decisions. And with the sort of all of the regulatory upheaval and the shift away from those mostly the large, centralized dealer groups to a lot more independence. And even a lot of the dealer groups now are really making it, giving their advisors more choice when it comes to tech. We’ve seen that ourselves. So, I think those two things, the concept of an ecosystem rather than all in one, but then having more independence at the business level to actually choose the technology that’s right for me, rather than having to make do with something that’s being dictated from further up the chain.

Rob (09:55.935)

So it’s industry dynamics that have changed significantly that has led to a bit more of an open architecture approach to how to build technology and hence you’ve really been right in you’ve been right in the right spot, I suppose, as that industry’s changed to give it independent practices, you know, self-licensed or otherwise, the opportunity to make their own decisions. But Just going back to the Fin365 development, two things happened at once for you that turned an internal tool for you in your company into something much more today. You were spending more than the planning business that you had your own yourself could fund. You were thinking, I can’t just s support this development on my own. So, you looked at other firms who were starting to look at it and go, hey, I like what you’re doing, Stephen. Can we actually get access? Can you actually make it available to us? Walk us through twenty sixteen, twenty seventeen, you did a small capital raise at the time, you kinda went full time and eventually sold your advice business. Tell me about that transition, about how you went from using an internal tool that was working for your needs to thinking, actually this is gonna be more than just for me.

Stephen (11:05.286)

Yeah, so that they were the two triggers that resulted in us spinning off the software into its own entity. It was the fact that for it to continue to deliver real value and real just not become something that sort of was static and didn’t develop over time like software can, we needed to spend a lot more money on it, and so that was trigger number one.

Trigger number two was we were starting to get interest from other firms that saw what we were capable of doing, especially around business reporting, business intelligence, workflow automation. And so that was what caused the decision to spin it off. I had a business partner in the advice business that probably thought I was a little bit crazy to try and start a software company, but he was also very supportive. So, I went full time onto the software. He continued to run the advice business for I think it was three to four years post that and we did a very small cap raise. I still remember the exact amount. I won’t say what it was, but it was just enough to fund that I wasn’t even paying myself, the Argent we had a couple of Argentinian developers that were helping us and it was enough just to lock them in and get them to commit for the ongoing development. And then we started pounding the pavement, looking for customers, setting up the website, and it was really it was bootstrapped in to a large degree and that was, yeah, twenty sixteen, so we’re ten years this year, and eventually got to the point where I knew I was never going back to the advice business. The software business had really overtaken it and so my business partner and I decided it made sense to just sell the advice business at that point. And he’s still in the industry; I actually talked to him yesterday and still a shareholder in the software business. So, it’s been quite a fascinating journey when I look back.

Rob (13:16.802)

Yeah, the founder’s journey. You weren’t paying yourself; you had a couple of Argentinian developers. You were bootstrapping it with whatever you could put together in terms of a small cap raise to keep things moving along. But now you’re roughly hundred and seventy-three firms across the country, in fact beyond Australia.

Stephen (13:33.659)

Hundred and seventy-four. Just another one today, Rob, since I last chatted to you.

Rob (13:38.369)

Okay, jeeze. I mean thinking about just the pace of growth, how many of you got queued up now? Because when I talk to people now, people say, we’re in the queue, it’s actually you know, because you can’t onboard, you know, all these firms all simultaneously, is there a bit of a pipeline that you’re keeping in terms of the new firms about to join Fin365 between now and Christmas, for example?

Stephen (14:00.251)

There is. There’s been a number of shifts in that regard. So, you always go through these waves of a lot of interest and then you eventually hit capacity and then you’ve got to go on higher and build out your own systems and your own ways of automating and improving the efficiency with which we respond to support tickets. I mean, it’s the parallels of some of the conversations we’re having with our customers around how they’re making their own pipeline, and client journey more efficient. We’re having exactly the same conversations internally. It’s just our customers are the, you know, the financial advice firms instead of the end client. But a lot of the same problems to solve. So yes, there’s onboarding you know generally takes anywhere from six to three months. We have a team of multiple people now that run that. That’s actually that actually runs pretty efficiently. So, there’s not really a waiting list, but we do the other thing that’s happened probably the last 12 to 18 months is just a sharpening the focus of the firms that we know are the most suited to Fin365, probably going to extract the most value out of the system, so that right at the beginning of the sales cycle we can have the informed conversations with potential customers so that anyone that comes on board is just a bit more aware of what they’re what they’re getting themselves in for. We still have, we still have questions about whether we have modeling, for example. So just making sure that the customers that are coming on board really understand it and that those that may not be as suited to it because and it’s not always the case, but they might be a smaller firm and they’re looking for more of that all-in-one system like the Pluto softs of the world are filling some of that market. So, we’ve added capacity and we’ve also got a lot more targeted to ensure that those are coming on are really ready to go and move through the onboarding process as quickly as possible.

Rob (16:17.134)

Yeah. So, between six weeks and three months to onboard a firm and you’ve now got two thousand users, which is you know, you’re actually as is as is evident from the conversations I’ve been having with others in the industry, you’re growing at a good pace now too, and you’ve got a reputation of being able to deliver on something which people are very familiar with. The Office three six five environment’s so comfortable and familiar to people, it’s a real plus that you’re bringing something that they don’t feel is kind of a new environment, new user experience to kind of figure out. What changed for you though at the point you made the decision to go from sort of startup speed, because obviously there’s a bit of a curve involved here when you’re building something, bootstrapping it and getting it going. What changed for you to go to sort of accelerate as a scale up business now and take on additional funding to really grow the business at a at a faster pace and make that acquisition that you’ve made to help deal with the support services around the software. Tell me a bit about that journey of from startup to scale up.

Stephen (17:21.989)

So, well it’s always driven by demand. And what we found was as we matured and developed that track record, in in the beginning, there’s always an uncertainty about any new entrants into the market when we eventually get to everyone’s favorite topic of AI. You know, we’ll talk a little bit more about that, but in the beginning, you know, when we launched, there were a three or four others that I can think of off the top of my head that launched around the same time that are no longer here. and so as we sort of got to year three, five, seven, and built a track record of delivering and also started to attract larger firms that generally have more complex needs, they’re generally more demanding because they don’t get to the size that they’re at without having thought through what their operations look like, what their workflows look like, how they are going to systemize, smaller firms tend to come in and you ask them what’s their client onboarding experience and they sometimes they can’t answer that question.

Whereas if I go to, you know, a company like a Stanford Brown and I say, What’s your client onboarding experience? They’ll pull out a visio diagram and say that’s what it is and that’s what the system needs to do. So very different level of maturity in those some of those larger firms that we started to attract. And then at that point, once we sort of got through that phase, it did kind of open up the floodgates more because we had demonstrated that we could deliver. So that was on always on the software side and everything leads from the software. But one of the other things that we discovered as we started to look at this scale up was the software, any piece of software might get you sixty, seventy, eighty percent of what you need. And we’re you know ten to fifteen years of development on the CRM now, so it’s pretty rare that someone new comes along and says, Hey, I need to store the favorite pet field for my clients, and we haven’t heard that before. And and so the core CRM largely meets the bulk of our customers’ needs. But there’s always some unique needs that our customers have, whether that’s you on the workflow side, Power BI business intelligence reporting side, they might have other divisions. This is where the mixed discipline started to flow in for us.

So, one of our customers does member benefits for corporates. we have other customers that do accounting and lending and and so they would come say well, you’re covering the financial planning, but we’ve got these other unique needs. We want to build on top of the Fin365 version, we want to customize further. And so, long story short, we discovered a gap in the market where the traditional tech consulting firms that you might think of going to help you with those custom needs didn’t work for our customer segment. And so, we tried that in the beginning. We say, no, we just do software. You go and over here, we can recommend a few people, go and talk to them about your custom needs. And time and time again, that it underwhelmed, it underdelivered. And so, we started very hesitantly providing some additional services, but the demand quickly grew to the point where we just didn’t have, we were borrowing from the software side of the business to deliver some of these additional services. And it was robbing Peter to PayPal. And we had started to work with company called Quirk out of the Philippines for our own internal needs and also some of those tech services resources, things like data migration, we were recruiting staff through them. And about after about four years of working with them, we sat down with their director Joe Hall. And said, look, this is growing really quickly, but we don’t have dedicated leadership. We haven’t put a proper business plan in place for this. It was just sort of organic. We really think we need to bring these two businesses together formally and then recruit and make the tech services side a real focus, which we did. So, we acquired Quirk in twenty last year, twenty twenty five, halfway through the year and it’s growing faster than the software at this point. The software’s still growing quickly but we’re coming off a slightly lower base on the services side and so yeah, it’s just accelerating.

Rob (22:15.404)

Yeah, because it’s so critical that piece, isn’t it? that you’re actually not just providing a software solution, which, yeah, as you say there, has really now going is going to serve most customers’ needs. You’ve been there around long enough to sort of have dealt with most scenarios. So, anyone out of the box can actually take on Fin365 and it does almost all of what they would want. But the services piece is you say there’s always elements of what people want to configure themselves that actually is their personal way of delivering. Whether it’s workflow or of something of that nature. So hence the services team you have that you acquired through Quirk has actually really obviously given you a great opportunity to support those firms. And it’s a very you know it’s a great thing that you’ve actually built that because I think many people, as you rightly know and everyone would speak to, they get very little value out of the software if they don’t put time into it. So often people will take software on and think it’s not it doesn’t do what I need it to do, but they really haven’t invested the time and effort to kind of really get the best out of it. So having a services team that can help guide them who has got experience at helping others do it, I’m sure is a is a massive value add to your client group to get the best out of the software.

But I want to touch on something that I was really interested in when I was investigating a bit more about your business and you went about solving the data feed issue because it’s actually not a small mass of that and I want to, I want you to talk that one through with us. But there were aggregators in the market when you started that actually kind of like pull data from the platforms and into one central HUB24 and you were then gonna look it to plugging into those to sort of they were the middleman if you like. You chose to hook up instead every feed one by one to each platform. Why did you take that harder road? And what did it cost you over that journey? I’m really keen to dig into this. And we had a bit of a chat about it when we spoke previously. It’s an interesting one because this is the area I suppose that many people think, well, you can’t replicate that. It’s just too hard to go one to one and get all these platforms to connect to you directly, but tell me about that.

Stephen (24:32.687)

What did it cost me? There’s definitely a few more grey hairs because of conversations with product providers. There were aggregators that we could have gone with. It would have been a soft option. I think this is where two thoughts. One was just my engineering background, which says anytime there are more moving parts, there’s more complexity and cost. And so adding an aggregator in the middle, they wanted to get paid.

Rob (24:36.568)

Ha ha

Stephen (25:00.179)

If something goes wrong, whose fault is it? Is it Fin365’s got a bug or the aggregator or the product provider? So, I knew that was going to be not ideal over the long term. But at the time it would have been the soft option to switch on and say, hey, we’ve got all these feeds, the other process for me was or the other thought process was really around data ownership. So, one of the I think different points of differentiation with Fin365 is we install into each customer’s Microsoft tenant. We do not host as the Microsoft Dynamics CRM in our environment on our customers’ behalf. And it might not sound like a big difference, but when it comes to data ownership and controlling exactly who has access to that data, it in my opinion, that’s one of the fundamental and foundational reasons that that financial advice firms are better off now than they were, say, ten to fifteen years ago, because they can own their own data. And when you think about your business and the valuation of your business, whether that’s I’m about to sell it and I need to show the buyer exactly what they’re getting and so I can show them all of this business intelligence and really, they can look under the hood and see all the data. Whether it’s looking for inefficiencies in my workflows and so measuring every step of the way and then having that feedback mechanism to make my business more efficient. More efficient means lowering the cost of advice. It means more clients per advisor, which ultimately means a more profitable business. So for me, owning the data and it and making sure the data was as rich as possible was a core selling point. So, it and it was the reason that we spun it off is I knew I needed data feeds for my own business. And it and the engineering cost and complexity of trying to take data feeds from different sources and pipe them into different destinations, not a single destination. It’s as far as I can tell, it’s a unique engineering challenge, at least in our industry. So yeah, it was going to cost me more and it was going to require some patience.

But it really is a key point of differentiation, and a real value add to our customers. We did make some different decisions, so I didn’t want, I wanted a clear line of differentiation between the CRM and the platform. It never really made sense to me that the CRM would try and overlap the platform and take in transactional data and do CGT calculations and all those sorts of things that the platforms are built to do. IPS does do some of that. We deliberately drew a line at investment holdings, asset allocations, unit prices. So, the portfolio report can be generated, but things like internal rates of return, CGT, historical CGTs, etc., most our customers would use the platform for that particular function. So that was a difference, but with regard to the direct connections, we’ve now got all the major wealth management platforms, there’s still a couple of legacy ones that we’re waiting on the back end hookup, but that’s in the pipeline, the in the beginning it was slow, it was really slow going because you’re a small startup, you’ve got ten customers, the big platforms don’t really want to waste time and energy talking to you when they’ve got their own priorities to focus on. Some were, you know, some of the timing was good. Ten years ago, HUB24 and Netwealth were still they were still trying to also grow. They were a lot more advanced than we were, but they were still in that nimble mode. And so, some of the hookups happened really quickly. And they were really happy just to hook up and they saw that as a point of difference for them to be connected to more than just Xplan. Some of the others have taken longer, but now that we’re at the scale that we are, we’ve got sufficiently large customer base, that the overall farm that’s managed by our customers on any platform is pretty decent. And so, when our customers collectively pick up the phone and call the providers, we get more attention than we did five to ten years ago.

Rob (29:41.913)

Yeah, that’s a really interesting and fascinating insight into the fact that the maybe the up-and-coming platforms that were nimbler and more flexible, they were prepared to back you earlier. More established platforms might be going, well, come back to us when you’re substantial enough for us to talk to you and but you obviously have you said you needed patience, but it sounds like you needed persistence as well because you could have easily sort of given up on that challenge, but the persistence has paid off. You’ve got those direct feeds. And yeah, I can imagine it was an interesting exercise trying to get big platforms to to acknowledge your existence and and make those one-to-one direct data feeds available. You just said there that you’ve you’ve made a very conscious choice about what you didn’t want to do, and you commented about IPS being Xplans, portfolio, modeling and accounting software. So, you’re essentially grabbing balances on a daily basis from platforms. So, you actually can deliver a balance sheet if you like, or a portfolio statement, but at the at the valuation level, not at that accounting data feed, sort of unit price level. Is that what is that what you’re describing?

Stephen (30:52.367)

Correct. And you never say never, but at this point we’ll never do transactional data. If you ask me today, doesn’t mean I reserve the right to change my mind. But one of the things we are seeing, because when you think about what the transactional data delivers, it delivers the ability to do things like internal rates of return, CGT calculations, etc. Some of the more

Rob (30:59.64)

Yep. Yeah.

Stephen (31:16.887)

Modern platforms are now expanding their APIs to also allow us to query for that those those data points without needing the transactional data. So I see the value of a customer being able to push a button and generate a portfolio report that’s that shows the the client exactly what the internal rate of return has been. I just don’t think we’re going to need to go down the pathway of transactional data because I think to continue to compete and be at a and differentiate themselves, the platforms are going to enrich their APIs and their data feed connections so that we can just query that information rather than needing all the underlying transaction data.

Rob (32:01.624)

Yeah, do you f reflect on that experience thinking that at the time I mean obviously now you’re looking back on saying we made the right decision, but at the time you were having to push with the platforms to get them to open up, did you think maybe I should have taken this in a two-step process, go with the aggregator first, get enough volume and then go direct after you’d built some scale through an aggregator? Or were you very happy now and on in hindsight to say, yes, it took longer, yes, it was harder, but I absolutely wouldn’t change a thing.

Stephen (32:29.487)

Yeah, I wouldn’t, I wouldn’t have done it differently. It’s the if we hadn’t of if we hadn’t have got any yeses, you know, let’s say we got two years into it and no one was allowing us to turn on data feeds, then that might have caused me to have you know, second thoughts. But we did get a couple really early on.

Rob (32:59.074)

Yeah.

Stephen (32:59.615)

And and and I was convinced, and it was a point I would make to any product provider that would listen, that ultimately their distribution is through the advice, the advisor channel. And if they were all things being equal, fees, investment menus, those sorts of things, administrative capabilities, all things being equal, if I’m an advisor, I’m going to choose the platform that makes it as easy as possible for me to do business because the cost of delivering advice is high. And so anywhere that I can find efficiency, that’s a benefit. So Netwealth and HUB24 were, I think they were two of the first ones that we turned on on the wealth management side. So now any customers that were using Netwealth and HUB24, it was more efficient for them to use those platforms through Fin365 than it was some of the legacy platforms that we didn’t have feeds for. And I saw it firsthand where because certain platforms were making it easier, advice firms just started to shift and say, okay, we can’t, we can’t take all of the money in one fell swoop. We know we’ve got a best interest due to all that sort of thing, but all new money can flow. And I think that’s really been the driver is we just have to be patient because,

Rob (34:08.718)

Yeah.

Stephen (34:24.887)

Integration between the advisor desktop technology and the platform has always made sense and the platforms that make it easy will have, will be the ones that will get the inflows. So that was the philosophy and and I think it’s it’s proven to be true.

Rob (34:38.391)

Yeah.

Rob (34:42.69)

Not only a smart move by you, but a smart move by them too, to actually see the potential in what you’re describing is, totally advisors listening to that will think, yeah, like, you know, we want to make it easy to deliver advice. We want the data feeds and if the if these platforms are the ones that’ll do it with our chosen software, Fin365, then new money can go there and perhaps it’s not just a small coincidence. I’m sure there’s other factors as well, but I keep hearing that Netwealth and HUB24 are getting the biggest flows across the industry. So, they took a more progressive position and they’re benefiting greatly from their foresight. So yeah, well done you and them. Let me just go back to that point. Well, I’m sure I’m sure it contributed.

Stephen (35:13.017)

Yeah. Yeah, I wish I could say it was all because of me they were getting the inflows, but the philosophy, we shared the same philosophy, so yeah. Mm-hmm.

Rob (35:32.439)

Yeah, and so they applied that philosophy in all aspects of their work, and so hence they’ve benefited from the flows they’re now receiving. I want to go back to that point you made about that foundational aspect, that really crucial thing about owning your own data. What should a firm ask a software vendor? Because you’ve answered this question already, but I want I want you to kind of be an objective observer here, what should a firm ask a vendor to work out whether they own their own data, or they just have access to it? What’s the question they’re trying to pose there to make sure they understand whether they own their own data or not?

Stephen (36:14.939)

Well it’ll somewhat depend on the system, but I’ll start with CRM and I’ll give you a real example and I won’t name them. But there was one of our competitors in the early stages. They were a hosted CRM. So that was you were logging into their environment, their tenant to use their system, they went out of business and advisors got locked out from their data. They went into administration. Administrator took over and I had customers, not even customers, I had potential customers and friends that I knew in the industry that were calling me saying, Do you have any idea how, is there any way I can get my data out of this system? I’m like, No, sorry.

So that’s if you are using a CRM where you are logging into someone else’s tenant and you don’t have full administrative control over that and I don’t just mean the Xplan system admin stuff. I mean full global administrative control to the point where if our customers wanted to or were new enough to be dangerous, they could press a button and blow up the database. You know, you can’t do that in any other environment. You don’t generally get that level of access. I cannot stop our customers exporting their data out of Fin365. That’s just not something that I can turn off. So, from a CRM perspective, that to me was data ownership. If I leave the dealer group, I don’t have to ask permission for my data. If I want to migrate for whatever reason Fin365 stops meeting my needs and I need to move, I will need to migrate my data, but I can get access to it because I have someone in my business has global administrative privileges in the Microsoft tenant. That to me is data ownership on the CRM side. The questions that you should ask other vendors is less about data ownership and more about how you are using my data to make sure it’s protected. And this becomes especially important in this new age of these AI tools that are doing very different things than what we’ve seen other systems do in the past. But yeah, pure data ownership, can they stop you from getting it if you decide you want to make a change?

Rob (38:41.826)

Yeah. I think the if nothing else, this episode I’m better there’s plenty, but that that one point you’ve just made there I think will make everyone who’s listening stop and think and ask the question because it’s such a crucial point and how scary it must have been for those firms who actually had data on a platform they had no can software they had no access to after the platform went into administration. Yeah, it’s not a small issue at all because our business is data. That’s essentially what we what we operate on.

Now I want to turn our attention to something that everyone’s trying to figure out. Advice production is the piece everyone is trying to devote attention and time to, to streamline how they deliver advice, because it is actually one of the the big choke points of the way the advice industry operates.

Stephen (39:11.8)

It is. Yep.

Rob (39:34.103)

It’s become pretty crowded and confusing, I must say. I mean, you’re actually on the inside of the technology and you would no doubt be hearing from a number of your clients that use Fin365 about different software they’re now trying to use as well as a complement to what you do. Map out what you’re seeing for us and what parts does Fin365 not play in and just give me some commentary a little bit about what you’re seeing with Claris and Paradino and there’s a number of them. Marlou, obviously I just was speaking to yeah, a whole host of them. Just give me your perspective at the moment on how that’s all becoming quite crowded, and how is it impacting the way your clients who are using Fin365, what are they saying to you?

Stephen (40:05.231)

Yep. Marlu, plan AI, if RetireMap. Yeah.

Stephen (40:25.787)

So, the first thing I’ll do is probably for the only time in this podcast, I’m gonna disagree with slightly with one thing you said. Well not completely. It is a choke point, but one of the things that when we have this conversation and we’re looking, because it always boils down to, I’m looking for more efficiency in my business. And The advice production piece historically has been a one of the most complex documents, or the most complex document that we have to prepare. The amount of data that goes into that one document is ridiculous. And even in systems like Xplan, fully full automation of that SOA was always a challenge. And you had to invest some significant time and energy. I think we’re going to talk about our friend Matt here shortly. He but he built an entire business about helping people automate the production of statements of advice within Xplan because it wasn’t simple. I do push back though when people are trying to figure out their technology strategy is if you look at the volume of SOAs that any business produces.

Rob (41:24.376)

Yeah. Yep.

Stephen (41:49.4)

Versus the volume of client reviews or the volume of annual engagement obligations or just other workflows that happen on a day-to-day basis in the back office. Generally, you can find more opportunities for efficiency improvements with simpler changes to technology elsewhere in the business. And so I will answer the question, but at the same time, it’s the I’ve always said the full automation of statement of advice to me was never the be all and end all of what I was trying to build with my technology. And you can just do the math. If it takes you six hours today to produce the statement of advice and through partial automation, you could pull that down to four hours. Well, that’s still a 33% improvement. But over here, if I’m doing, you know, if I’ve got 300 existing clients or 150 per advisor, and therefore I’m doing three client reviews per week, but I’m only doing two SOAs per month, then there’s probably more opportunity for efficiency gains on the client review site.

So, when we have these conversations, we tend to try and l take it up a level and not just focus purely on the SOA. With that said, you’re right, it’s really crowded, and one of the decisions, not in a similar way to we made certain decisions around data feeds, we made certain decisions around advice production where we said we are not going to try and be an all-in-one solution. Now we do have a document templating system. We have other document automation systems that we integrate with, but we don’t do modeling. We don’t do product research. We’re not trying to compete with a ProductRex or a WealthSolver or an Omnium. Rather, we have integrated with other systems over time. So, we actually do have an integration with Xplan through the open AI, the Xplan open AI.

Stephen (44:06.041)

A number of our customers still do use Xplan for the production of the SOA, but it’s also fair to say that the majority are looking for alternatives. And I think they’re really looking for for two reasons. One, the Xplan modeling, while it’s quite powerful, was never really a great client experience. So a lot of our customers are looking for projection tools, that they can use also as part of the client presentation piece but also just the time, the ease of use. Because even in Xplan with automation, you were still spending six, eight, ten, twelve hours, depending upon that complexity of the statement of advice. How do I get that down to two hours? I had one customer at one point that said, why can’t SOAs just be 20 minutes? Why haven’t we solved that problem? Like I’ll probably never solve that problem for you, but that’s what advisors would like to see. How do we sort through the noise? we the Quirk team has been useful here, our acquisition of Quirk, because as well as the technology services, they still do traditional paraplanning outsourcing. They do client services, client admin outsourcing. So, a lot of our customers, and and Quirk has some of their own customers, still use those traditional back-office services. So, Quirk’s got an entire paraplanning team and therefore is getting exposure to all of these new tools as they come online with certain customers. And there’s actually a webinar, I think it’s in a few weeks, where they’re going to run through a comparison, a side-by-side comparison of many of the tools. I don’t know. I think at least Marlu, Plan AI, and if are definitely in there. RetireMap, I think, is in there, but we’ve probably got about 15 tools at the moment that We’re being asked to integrate with not all advice production. Many of those are just some of these new AI tools that do simpler things, but yeah, it’s pretty crowded.

What do I need to so without picking a and or endorsing any because they’re moving so quickly and I don’t do advice anymore, so I’m really not the bet one who’s best position. I’ll comment on the integration piece. So, at the moment we have integrations with IF, who has just built one, and so we’ve worked very closely with them. And there’s a number of our customers who have started to use if and are really happy with the results. I know that when they do a comparison with, say, X Tools Plus, one of the things that’s been challenging for all these tools is the complexities of the Australian tax calculations for SMSF and Supra and Family Trusts, etc. And that takes longer to build. But in general, the feedback I’ve had is that the if advice production tool, which does hook into product recs, does hook into Omnium, will do a document, is delivering some great results, albeit early results. RetireMap has focused more just on the modeling, but they do the complex tax calculations. And the thing I like about both of those tools, and we’ll get onto AI shortly, they are both deterministic modeling tools. They will give you the exact same result every time. And for me, when it comes to the accuracy of modeling and the actual deli advice on delivering, I do have questions in my mind around how the AI tools are going to attack it and whether they’re gonna take a different approach and what they can achieve. I don’t have a good answer for that yet, but it’s a question, it’s a question in my mind.

Rob (48:02.742)

Yeah, it’s a critical point, isn’t it, that you need that software to be deterministic because you can’t get variable results from the same inputs. Whereas AI really is probabilistic and therefore you will get various results for the same query. So, on the back end of doing your modelling it must be deterministic. And as you say, RetireMap, who we had Andrew Gardner and Ivan Gower on as guests talking about their software

Stephen (48:26.661)

Yep.

Rob (48:30.22)

And making that point very strongly that it that’s their part of the ecosystem. And as you pointed out there, they are one of the providers in the market that does modeling at the level of sophistication and depth that perhaps everyone’s only been used to getting out of X Tools Plus in the past, whereas many other modelling software didn’t have the complexity, didn’t go to SMSFs because it’s kind of exclusively to Australia, perhaps didn’t tackle family trusts but having spoken to Anthony War Bianca Musiko from War Hunt on this podcast as well. They spoke about that when I spoke to them in Melbourne about six weeks ago now. Had caught up with them for a bit of morning tea and they said that yeah, they run a sophisticated business. So, I thought if they’re using RetireMap for their modelling and they use you, Fin365 for their CRM, then they must be satisfied with the level of depth of modelling that they can do. But I wanna turn back to another guest of the,

Stephen (49:18.907)

Correct.

Rob (49:27.576)

Podcast thus far, which is Matt McGuinty. You flagged that we chatted about Matt there briefly. Matt was episode thirty-three guest, he was at the time running his own business on, as you say, advice production integrating with Xplan. Full disclosure, we use that software still, he sold his business to the Umlat team. What did Matt’s decision to sell his business and join you?

Stephen (49:29.371)

Mm-hmm.

Rob (49:56.8)

Signal to you and what is he doing for you now?

Stephen (50:02.155)

Well, the signal I think was when Matt reached out, he just gave us some feedback from an outsider perspective, and you touched on it up front. You know, you’ve it it’s a small industry, it’s actually not that large. People tend to know each other then you know, as you said, Warhunt uses us, RetireMap integrates with us. When Matt reached out, and I didn’t know Matt before the initial call, probably because he lives in Perth and spends all his time over there. He just made some comments around sort of what he was hearing and not so much about just the pure capabilities, but more about our approach and the ecosystem decisions that we made and the philosophy there. He’d picked up on the services that we were now offering, the tech services side of things. He was aware of the quirk acquisition. So, it was interesting to get some outsider feedback because when you’re in your day-to-day bubble and you’re dealing with everything that’s on your plate, and it’s not always perfect and it’s not always smooth sailing, there can be moments where you lose sight of the progress you’re making or,

Rob (51:25.848)

Yeah.

Stephen (51:26.099)

Or whether you’re delivering on what the customers really need. So, it’s always nice to get some external feedback, well, at least when it’s positive, and so yeah, he sort of indicated that there was interest in working with us and we actually were recruiting. So, we talked about the acquisition of Quirk. Once we finished that, we knew we needed to bring in some senior leadership on the tech services side, both on the consulting and the delivery, because they are really the two, the two services, sort of top level first points of contact. Customers come to you and say, I want to get more out of the system and often that’s a consulting exercise because they’re not exactly sure what they’re looking for. So, you have to unpack that and that’s what Matt’s doing. So, he’s the first point of call anytime customers come in and say, look, we think we might want to extend or customize or configure or build a new workflow. We need better reporting. He’ll sit down and unpack what that looks like. So that’s his role and then we recruited another person to handle delivery who’d had 20 years working in dynamics consulting businesses, building offshore teams. So, Matt and Rosie is her name came in at the same time, and we’ve largely sort of handed over the running of the tech services business to them.

Rob (52:52.91)

Yeah, so you’ve got the software off the shelf, you’ve got the consulting part of that now to help people make the right decisions as to how they implement the software. You’ve got the tech services team to kind of execute on that, design the way that Matt’s consulting with those firms. So, yeah, really great experience onboarding, I’m sure, for firms that are joining Fin365, and I wanna just touch on this tech services team a little bit further. You said you’ve drawn a hard line that you won’t sell tech services to firms that aren’t Fin365 customers. I think you said their Quirk had their own customers and perhaps they’ve got some legacy there. But you aren’t trying to use this as a standalone tech services team that can support anyone under the sun who wants to get support on their technology. Really, it is a solution for Fin365 customs exclusively, leaving revenue on the table perhaps. Why do you think that’s the right call for you?

Stephen (53:43.867)

Mm-hmm. This kind of goes back to earlier in the piece where I said we’ve just become more focused, there is the reason I think we have had the success that we have over the last 10 years is because when you look at the big end of town when it comes to technologies, so your Microsofts and your Amazons and your Googles and those companies. They build the broad platform, they build the enterprise capability, but they’re generally not going to go after industry verticals. they may dabble like the Salesforce Financial Services Cloud and things like that. But they’re never going to go laser focused onto an Australian financial planning industry vertical. They’re definitely not going to do it for the small to medium corporate. So they might go out and do some work for Commonwealth Bank and build Some custom work for them. But we have we have chosen such a unique niche, Australian financial planning and financial services, small to medium, that both on software and services, the larger organizations and the ones that can really deliver what our customers need. They’re not interested in servicing customers that are fifty, a hundred, even two hundred staff. Mm is not going to be show up on Microsoft’s sales radar. So, for that reason, we know that there’s a gap in the market. We happen to be uniquely positioned to fill that gap. And so, maintaining that focus means we’re not, spreading ourselves too thin and we’re past the point now of and we did this in the early stages of jumping and saying, yes, yes, yes, yes. I’ve got a, you know, I’ve got in the early stages I had a small wealth management platform that was white labeling Macquarie who needed a CRM and this they’re actually still a little customer. But we wouldn’t take on a customer like that anymore, because we’ve just, we’ve figured out our niche. We know we can deliver. And there’s more than enough to keep us busy.

Rob (56:10.146)

Yep, makes sense. Stay focused and go stay narrow and deep and look after that whole workflow from end to end. So, it totally makes sense to me too. I just thought it was worth asking the question. I want to talk APIs with you. You said there that you do hook into the Xplans open API. But you said the dynamics, Microsoft Dynamics API is a very different animal to Xplan’s version of that. What does what is the difference and why does it matter?

Stephen (56:38.085)

So, this this the answer to this question will flow into the you know the some of the conversations that are coming up on AI because there’s a question that I think one of the questions I think everyone needs to ask when they’re exploring AI tools at the moment is do, I actually need AI for this? And one of the things to understand about these different systems, especially when it comes to integration, is what’s possible through the more traditional mechanisms of integrating with other technologies, building out customizations, etc. The short way to answer your question is if I go and to be honest, I haven’t looked at the iris spec for a while, so it might have evolved slightly, but at the beginning, when I went and queried Xplan and said, I would like to pull in Desmond and Molly Jones, the client, please give me their data. Xplan would return Desmond and Molly Jones, and they would give me a certain set of fields that they had predetermined were ones that they would allow me to access for Desmond and Molly Jones. Dynamics API gives me full access to the database. So, I can go in and query dynamics, not just for the core Microsoft fields or the core Fin365 fields. If Asensier, who I know Matt’s coming on in the not-too-distant future, all of their customized yeah, different Matt, all of their customizations, they’re currently building an integration themselves, some AI capability that’s hooking into dynamics through the API. It can query for their customers,

Rob (58:18.19)

Is Matt Fenning, yep, Censia.

Stephen (58:34.947)

Fields, their customizations. Whereas that’s not possible through the iris. And one of the one of the challenges with the integrations with Xplan is you’ve got so many different versions of Xplan around the industry that it’s you can’t build a fully rich and functional integration because of the limitations of that API. That’s not the case with Dynamics. It’s fully open. It’s Microsoft documentation on their website. So, what you can do in terms of read and write and query and permissions setting is it’s just a much richer capability.

Rob (59:19.374)

That makes sense to me. So now in another area here that has captured a lot of attention recently. This model context protocol capability. Demystify MCP for our audience because there are some fairly strong and contradictory opinions floating around of API, which is you know that the way you just described API, sort of there’s a known data point you can access that data point, you know, interrogate the system and pull data from one system to another. Is MCP a different kind of pipe? Or is it just an interface layer sitting in front of those same APIs you’re describing?

Stephen (01:00:01.538)

It’s the latter. So I’ve seen a number of LinkedIn posts and other commentary around you know MCP is sort of the new way of connecting, but it’s actually it is just a new layer on top of your traditional way of connecting. So an API has a predefined set of instructions. And for someone to build an integration, so when RetireMap or IF built their integrations with us, they needed to know the API, they need to know our data schema. And specific to the functions that they wanted to connect in, they had to know how to query the API. So for the modeling, they’ll say, please give me the household and I need the members of the household and I need the related entities. Tell me if there’s any trusts, tell me if there’s any SMSFs. I need all the assets and the debts and the incomes expenses. Every one of those is a predefined query. The MCP and an AI tool like an agent is not it doesn’t build to that API. So, for an agent to spin up quickly and to be able to interact with a database, what MCP does is on the back side of the MCP, it talks to the API. It knows how to query. And then on the front side, the part that the agent talks to, it has a set of actions or activities that the AI can understand. And it says that you can ask me about clients. And so, the AI doesn’t need to know what that means from an API perspective. It just says, please give me a list of clients for whom Stephen Hanley is the advisor. And it will ask the MC at the front end of the MCP, and the MCP has logic in there to say, okay, agents asking about retrieving clients. I did a query for the households and the contacts, and I’ll return that data. Now it’s making decision about what it returns. And so, what that means is you still not fully deterministic. And just because I say, give me a list of clients, depending upon how you ask that query, sometimes you might get the exact same response. Sometimes it might say, okay, well he didn’t really specify that he wanted active clients or inactive clients or clients for whom he was the advisor, or A versus B versus C clients. And so, it’ll make its own decisions. If I yeah, so that’s the difference. It’s the,

Rob (01:02:41.294)

Yeah.

Stephen (01:02:44.867)

You still are talking through an API. It’s just the MCP makes it possible for an agent or an AI system to access the database without having to understand that API.

Rob (01:02:57.71)

That’s helped me understand it too, and I now appreciate the precise nature of an API versus the MCP server connection, which is more of a, as you say, an agent asking a query of that API without necessarily being deterministic as to the output. You’ve recently built an investment analyst agent, and you said you built it in two days through MCP is the way you intended to actually have it operate. You demonstrated that it got the reaction you wanted, and then you found it produced a slightly different answer every time it ran. Talk us through what you’re doing with that now to build that deterministic back end, and an agent that does less of its own reasoning to get the output that you would expect every time.

Stephen (01:03:42.789)

Yeah. Yeah, so it’s still a work in progress. We’ll actually have an AI webinar in a couple of weeks where we’ll give a little sneak peek, I actually built it originally at for our Symphony conference in February. So how long? Six months ago, now. And it’s an indication of the idea of MCP. MCP prototype to something that’s productizable at scale. It can take some time. So, in the beginning, because you get up at a conference and you want to show people what’s coming, what’s possible, I spent about a month before the conference just exploring what the latest capabilities in Copilot Studio were. And I bought built about six agents over the course of that month. And I just it was very simple. So, the MCP makes it really simple. And I just gave the agent a set of instructions I gave it a template. I said, okay, this is what an investment portfolio report looks like. I need you to pull in the assets, pull in the holdings, compare against the model portfolios, look to see if there’s any drift and any of the asset allocations are outside the model. We hook into Zenith for this particular agent. So we’re able to pull their research data in as well and compare, has there been any downgrades or anything like that? show me if if there’s been a downgrade, go and find me my clients. Who are affected by that downgrade. So that’s the purpose of the agent. And yeah, the MCP made it possible for me to spin this thing up and give a pretty compelling and interesting demo of what was possible. But every time I ran it, it would give a slightly different format, a slightly different outcome. And when you are generating reports for clients on their investment data, that’s never going to work.

So, we took it back and you know I have, I’m technical but I know my limits these days. They don’t let me put my hands on code anymore. So, but I took the concept back to the development team and said, okay, we need to, I need you to tell me the best way to build this and so, the develop the primary developer who’s working on it spent some time and he said, look, I actually don’t think we should just let the agent do it all. And so, the method that we’re using is a hybrid method where all of the querying of the information that needs to feed this report and actually the report itself is quite structured. And then the agent sits on top and just makes it easy for a user to request the outputs. So, it’s a more canned set of instructions, including some instructions back to Zenith. So, you can you know query just the Zenith research data independently. But when it comes to that investment portfolio report, all of the heavy lifting is being done by a deterministic engine that’s talking directly to the Dataverse API and then feeding the output of that to the agent to say, here’s and this is getting a bit technical, but here’s a large JSON file. This is how you build the report. And then the or it might be HTML, can’t remember. yeah, so we used a hybrid approach.

Rob (01:07:11.564)

Yeah, that makes a lot of sense. We’ve had a bit of a I guess a play around with a probabilistic AI to generate documents, but as you say, if you don’t have a deterministic back end, you’re going to get variation in the output based on, you know, even the same query. So a hybrid model where you use the MCP at the front end to sort of if you if you like curate your question set, get to the point where you’re being asked to completely answer all the questions that are logical to ask. But then when the query is sent through, it’s sent through, I guess, in a standardized way. So, the API is getting the answers it needs in a s in a standardized fashion, I guess. Is that being that have I described that well enough? Okay. Yep. I’ve kept you longer than I promised. We’ve gone out well past the hour now, Stephen. So, I appreciate I’ve got one more question if that’s okay.

Stephen (01:07:57.541)

Close enough. Yep. That’s okay. Sure. Yeah.

Rob (01:08:08.696)

There’s a trusted advisor out there listening right now that’s decided their business needs more than Xplan and they’re being pitched by fifteen vendors. We’ve already talked about the confusion o in the market. What is their first move and what should they deliberately not do yet?

Stephen (01:08:26.651)

So the move that I made, so this is the same question I asked myself fifteen years ago. I determined that the current tools that were available to me were not going to meet my needs. And therefore I was going to be looking for one or more alternatives to to to meet the needs, to to perform the functions I needed in my business. And it’s like building a house. You start from the ground up. You don’t go and start from the cool AI tool. That might be a great tool, but depends on the data that sits underneath it, the information that it have a has access to to perform the functions that it needs to perform. If you do not have the data foundation centralized, clean, complete, well organized, all of the relationships between the different data points are well structured. If that data source is also not connected into other sources of information, all of the files that you probably have sitting in SharePoint if you’re using the Microsoft environment. If all that, and then your permission structure around that. I still remember the the Microsoft story of an enterprise customer who wanted to roll out Copilot, and one of the first questions that was asked in a test was, What’s the what’s the salary of the CEO? And The agent answered the question because somewhere s sitting in a file somewhere in SharePoint without the necessary permissions was accessible to that user, and therefore the user was given the answer. And to this day, when we bring on a new customer and we go in and we analyze okay, what is your SharePoint structure look like? What’s your data structure? There’s always cleanup that’s necessary. So, you’ve got to start with the foundation and some of that might even come before any tool selection. So often we will have customers that are or potential customers that are really keen on what Fin365 can deliver, but there’s still a bit of work they need to do to get ready to be able to take advantage even of our CRM, which is a foundational piece.

So, while you are exploring alternatives with when it comes to technology, alternative systems, you can start to put a strategy in place. And the first step in that strategy is generally organizing the foundational layer, the data and the files, the information, while you are searching for the right tools to help with that layer. Once you’ve got the foundation in place, then modeling AI file notes, all of the functions, it’s client portals, all of the things that sit above the foundation. They start to light up because it doesn’t make sense to have a client portal if 50% of your clients are missing a date of birth, the spouse not connected, you f you haven’t marked one of the spouses deceased who died two years ago. You you don’t want clients logging into a client portal if the underlying data is not clean. So you’ve got to start with the foundation and then work your way up.

Rob (01:11:44.697)

Yeah, that’s great advice. I think anyone listening would it’s just it’s such a sound piece of advice because it’s easy to get distracted by what’s available in the marketplace now and start from a different position around what’s what can do planned faster, what can do file notes faster. And so looking at that foundational piece and where all the data’s stored. You talked about data ownership, which is a really key point that I really think every listener will will take heed of. And just that getting the data model right so it’s clean before you start thinking about how you’re going to use your data with third party tools outside of the CRM. I appreciate your time, Stephen. I always learned something talking to my guests on the podcast and today has been no exception. So thank you for taking the time, spent spending a bit of a a later evening talking to me over there in Texas. And I wish you continued success. good friends of mine are using your software and they’re reporting nothing but good things. So I appreciate you taking the time to fill me in on what’s happening at Fin365.

Stephen (01:12:43.717)

Thanks Rob, appreciate you having me on. Happy to come back anytime.

Rob (01:12:48.206)

Terrific. Thanks again, Stephen.

Stephen (01:12:49.372)

Cheers.

Scroll to Top