Join Heather Hopkins and Alasdair Walker at Handford Aitkenhead & Walker as they explore the transformative impact of AI on financial advisory firms. Discover how industry leaders are integrating AI into their practices and preparing for the future. In Episode 2, Heather and Alasdair dive deep into the world of AI in wealth management with Jonathan Stubbs from LIFT-Financial. They uncover how AI is transforming the industry today, explore its future potential, and discuss the hurdles of implementation.

Watch via:

Transcript:

Heather: Hey, welcome back to NextWealth Insights. My name is Heather Hopkins, Founder and Managing Director of NextWealth. I’m joined by my co host, Alasdair Walker Managing Director of Handford Aitkenhead & Walker Ltd,  Chartered Financial Planner. Hello, Alasdair,

Alasdair: Hi Heather. Great to be back again. We’re going to dig a little deeper into AI in episode two of this, which is the companion season for the AI Lab. I’m absolutely buzzing with plans and ideas after our event on the 17th. We wouldn’t be able to do any of this without the help of our four amazing sponsors as well, and they are Aviva, Fidelity, Salesforce and SS&C, they’re all really kind of invested in the development of AI, and I’ve always thought it’s the it’s the bigger players that we need to encourage to think deeply about this, because it’s going to help drive the rest of the sector forward. On that note, on the note of our AI Lab event, I’m really looking forward to chatting with our guest today. I think it’s fair to say he was our most popular speaker at the event last week. Over to you, Heather to introduce him.

Heather: Yeah, I totally agree. I’m buzzing after last week’s event, and the feedback has been phenomenal, particularly for Johnny’s session, which was a really, really practical one. So Jonny Stubbs is director at risk at LIFT-Financial Group. Jonny’s a chartered financial expert with 20 years in the industry, he leads LIFT groups, compliance and para planning teams, overseeing regulatory and technical responsibilities, fostering innovation and professional growth. And I can attest to that because I was at a paraplanners assembly event recently, and a bunch of his team were there, and they were not only saying really nice things about the AI tools using in the business, but also about Jonny. Jonny’s passion lies in merging cash flow based financial planning with a forward thinking embrace of technology to deliver effective solutions to drive LIFT’s continued progress. Jonny. Welcome to the podcast.

Jonny: Hi Heather. Hi Alasdair. Kind words, both. Thank you.

Alasdair: So when, when we interviewed our last guest at Peter from SS&C. Our first question to him was to sort of describe AI. And he’s a machine learning expert with a with a background in that side of things, academically, really and practically. And he gave a really good high level answer. So if people are looking for a high level answer to AI, check out episode one of this season. But I thought it might be quite interesting to ask you, as somebody who is, I think, a self professed AI geek, what AI means to you, not just at a high level, but kind of practically as a financial services practitioner.

Jonny: It’s a really good question. When we first started looking at AI, I would have probably answered it quite technically. It’s an LLM, you know, it’s uses natural language and generates text, blah, blah, blah, but now we’re sort of getting on two years down the road. I’m just seeing it as a tool. It’s just another tool that is helping our staff to work more productively and efficiently. And I think until AGI comes along, I think it’s just a tool,

Alasdair: Just to dive in with a with a glossary there. LLM being a large, a large language model, and AGI being the quantum computing of the AI world. I think it’s probably fair to say, and that’s a as an artificial general intelligence, you know? So it’s like, it’s that thing that’s always just five years away, but, but never, never, quite sooner.

Jonny: Yeah, that’s right, yeah, sorry, some of that tacky bit Does, does come back in,

Alasdair: I know the feeling

Heather: I always think of AGI as what we’re able to do now is we can train a robot to jump like a dog or play chess or write a suitability report. But what we can’t do is have it do all three and and that’s what AGI is, is it’s, it’s replicating human intelligence, which is intelligence across a number of domains. So that’s my simple way, my simple my simple brain, the way of explaining it.

Jonny: I was listening to it was, I’m catching up on some of the implement AI podcasts, and they were talking about open AI have come up with a four or five level stages of AI. And what they’re saying is they’re only at level two at the moment they think, which is basically AI being a tool. And I can’t remember where they put AGI, sort of artificial general intelligence on there, but it was a four or five, and I thought that was really interesting. And another take on it is that we’re, we’re going to have to redefine what we think of as intelligence, because we’re, if we’re after human intelligence, there’s so many AI tools now that are arguably bettering human intelligence in some aspects. So it’s, does the whole point of general intelligence even need rebasing? So I think there’s interesting conversations going on around that.

Alasdair: Although I think it’s fair to say anybody who uses Excel regularly knows that in very narrow domains, computers have been more intelligent than humans for an awful long time. The slowest part of any Excel spreadsheet is always the user, right?

Jonny: True. Yeah, Yeah.

Heather: So, Jonny, you shared at our AI Lab event that we recently had some of the journey that you’ve been on at LIFT with AI. Talk talk our listeners through, you know, what was the problem you were trying to solve? It’s, you know, these are fun toys, but, but you but you really shared some practical business applications. So, what’s the journey you’ve been on? Why did you go on that journey, and how’s it going?

Jonny: So our journey at LIFT began with my curiosity. Really, as Alasdair has said, I’m a bit of a nerd when it comes to tech. And in late 2022 I first started playing around with chatgpt after it started to go viral on social media. So over the next few months, more tools like Claude, Google, Bard, eventually quite a bit further down the line Microsoft copilot, along with image generators like Dali and mid journey started being released. And I was just having fun playing with these things, fun gadgets, getting them to write silly poems, etc, showing them off to, you know, friends and family, what have you. But I then started turning my mind to actually, is there something in these two to help from a work setting? And the obvious one that’s always been the bane of my life being a power planner and manager of power planners is suitability reports. So I just started playing around with chatgpt, asking it to generate suitability reports. And I think quite a lot of people, having spoke to them, have tried this, and you don’t get bad results with just quite a simple prompt. So that just got me thinking, look, if you actually did this properly, and not me just playing around at home, if you had a, you know, machine learning experts and AI experts doing this, and developers, where could you get to? So off the back of that, I introduced these tools to our senior leadership team, who, fair to say, weren’t quite as nerdy as me, and we all got subscriptions to chatgpt when it started being a paid for version. And yeah, they quickly got on board with it, not being as nerdy. They still saw the potential, you know, from a personal work efficiency point of view, using it to help them write emails more efficiently, communications, brainstorming, those kind of things. So once I’d got the sort of the core senior leadership team on board. We then approach the board who, it’s fair to say, I’ve always been progressive on tech. They’ve always been keen to use tech. We’ve, you know, we’ve, we’ve never struggled to get any projects through that need a new license or something like that. So we were hopeful that they’d get behind it, and they were, like, really excited when we told them where we thought it could go, what our ideas were. And they said, Yeah, just crack on. So we did a bit of sort of research, looking around what tools were out there. But the real breakthrough came after listening to the implement AI podcast with Piers Linney and Alex shuckler, and this was, it’s a really good podcast, if anyone isn’t aware of it, well worth a listen from the start and off the back of that, they set up a conference in London, and I attended that, and there was someone there who just mentioned this AI tool called advisory AI, who were looking, they were a startup, and they were looking for new customers, and they were trying to generate meeting notes for financial advisors. So I got in touch with Alan, who’s the CEO of that company, and, yeah, it was kind of we hit it off. He was really excited about the potential having been the ex IFA himself and we started working together. Fast forward to today, and our advisors and power planners are using AI tools from advisory AI daily to help them work more efficiently.

Alasdair: I think that move from sort of experimenting and trying it out, and you’re obviously an early adopter of that, probably long before anyone was panicking about privacy policies and GDPR, but we’ll brush over that, going from there to going on a journey to bring not just the senior leadership, but also the team around you, with you is a bit that I find particularly interesting. I can’t remember who mentioned it at the event, but somebody talked about this concept that for every pound you spend on the tech, you’ll need to spend five on managing the change to use the tech. So I do feel like what you’ve told us. So far is that kind of, you know, step one, will do something. Step two, question mark. Step three, profit. So like, what happened with that? That driving the change? How did you how did you get from this is a great idea. You’ve persuaded the board to actually getting people on the ground to use it. I suspect that’s what scares a lot of our listeners.

Jonny: Yes, if we take it, because there’s two aspects of this, there’s, two main projects we’ve been working on, meeting notes and suitability reports. We started with the meeting notes because they were the lower hanging fruit, slightly easier to develop something just because there’s more nuance involved in suitability letter generation. So it’s fair to say that it’s been a journey to get everyone on board. And it wasn’t easy. However, we did it in stages. I think fair to say, the advisors probably get that in our firm, and the paraplanners get that I’m a bit of a tech nerd, so first of all, it wasn’t unexpected that I’d be coming to them with something along these lines. And we’ve got an academy. We run an advisor Academy. So quite a lot of our advisors are younger, arguably more tech savvy. They were, you know, starting to get interested in AI themselves, you know, after me sharing stories about chatgpt and the like. So when I first proposed this and gave them the pitch for it, and they were really excited. So what we did, we brought on board those advisors who were really keen on the proviso that, look, initially, there’s probably not going to be a good return on investment from your time. It’s going to require more of your time input to develop this, to get it to a point where it’s saving time, and they got it. And to be honest, that’s probably the single biggest thing that allowed us to, well, allowed advisory AI, Alan and his team to develop it so quickly, because we were providing lots of regular feedback and it was an iterative process then, and the tool just got better and better until it, to be fair, it wasn’t. It was probably weeks, couple of months, before we’re in a net positive position, I’d say, with time saving on the meeting note.

Alasdair: And you just touched on a really interesting point there as well, which you which you made a point of at the event and I think it was a bit of a light bulb moment, certainly for me, and I think Heather might echo that about that kind of return, and looking for a return on your kind of AI investment, and not just being happy that you’re driving it forward, but actually demanding some some returns. Can you talk to that a little bit for us.

Jonny: Yeah, yeah, it’s well, obviously, you know, as a senior leadership team. You’re always working on projects, but to be able to prove that they’re successful is not always easy, especially when the some of them are a bit wishy washy, which they can be in financial services. Now, one thing I think we benefit from at LIFT is we’ve always time recorded. And I don’t mean that in the sense that we’ve time charge on an hourly basis, but it was driven by our FD, actually, from an accounting perspective, and a just sort of a management oversight point of view, having that kind of data how long things take in your business is just been absolutely invaluable to us as an SLT over the years, and this was another area we knew how long things take, from meeting notes to suitability reports. So it was easy for us then to be able to see whether there was a real return on investment with this. Plus, you know, anecdotally, it’s quite an easy one to prove if an advisor is coming to you and telling you, God, it’s taken me, you know, 20 minutes to do a meeting. No, it used to take me two hours. This is Game Change, and I love it. Can I have more? We immediately knew we were getting a return on investment.

Heather: It’s really interesting that journey of persuasion that you’ve been on because you started talking about that you and your curiosity, talking to the exec team, presenting it to the board, getting those younger, keen, you know, advisors who are sort of interested in tech, as early adopters, who become advocates that then help to persuade others, have you. I mean, have that? People always say it’s the people that are the hardest. Are they? Are they obligated to use it? Is it optional? Has there been any, you know, any resistance to adopting the tech?

Jonny: On the meeting notes side, I think initially, some advisors were skeptical, and I don’t blame them. When, I was doing my pitch, it sounded too good to be true. So, and this was the point, I think, where we are now with AI, everyone’s accepted that it can add value and it can drive efficiencies but the point we were bringing this in, it was just a buzzword. It was just, is it another one of these tech fads that’s just going to disappear? And when I was saying, Look, this is this could make your meeting notes take 20 minutes, as opposed to two hours. Are like, Yeah, whatever. You can understand why they were skeptical. But then, like you say, once we had those first few advocates, what I think helped get everyone on board was advisors talking and, you know, sharing their examples with the other advisors. So the ones who were a bit reluctant when they saw that others were using it. I suppose it’s that, you know, you don’t want to get left behind. So there were a few, just because of the quirky, some of our advisors work in slightly different ways within the business. They were later adopters. But everyone now is using it in some form or another, and we’re just trying to sort of increase the amount they use in it when it came to the suitability report project. So this is the meeting notes. Are they involve the advisors. That’s benefiting the advisors saving time, but the suitability report that’s obviously going to benefit the paraplanners make them work more efficiently. So again, understandably, there was some reluctance there, I think, to adopt something that, in theory, if you were reading all the news stories, oh, AI is going to take our jobs. So I think there was a little bit of that. I’d like to think I explained why. I didn’t think that was the case. Obviously, you know, we’re not looking to introduce AI to make all our paraplanners redundant. Because, frankly, as good as it is, it’s still not as good as a well qualified paraplanner. There’s it just can’t get that nuance. You’ll always need that human input at least. You know, maybe 30 years in the future, maybe, but right now, it’s just a tool. But what I said to the paraplanners was, look, what I think this will do in future is make your jobs more interesting. It’ll take away some of the mundane aspects of the job. You know that that getting a Word document up, filling in the client’s name, policy, information, values and also what it does. I don’t know about you guys, but I’m not a natural writer. When I was a paraplanner, I really struggle getting that first draft of something. Once I’ve got that, it’s easy, but that first draft always took me ages. And this is the beauty of what the generative AI can do with on the suitability letter side, it gives the power plans a first draft of objectives, recommendations, disadvantages, any other sort of free text blurb you want in your report that’s contextually and accurately linked back to the meeting that took place with the advisor. So that’s our that’s where we see the main drivers of sort of time saving on the suitability report is that pre populating the suitability report and doing that first draft, the paraplanner will then come in and top and tail it and just, you know, add their experience and knowledge and human skills.

Alasdair: One of the interesting things that one of the other speakers, Professor Adrian Hopgood, said, that really stuck with me was about not just accuracy levels, but also the concept of false positives and false negatives. And he was talking about in a medical domain, and false positives and positives. False negatives, you know, could be life or death, but I think about the kind of the suitability report side of things as a financial planner, and I think those are your high stakes documents. So not quite life or death, but could be life or death of the business. And I kind of think I would be really interested to hear, from your perspective, what sort of metrics you’re measuring there in terms of not just accuracy, but also is the AI model missing things that it needs to not miss, or creating things that it needs to not create? And how do you get a sort of positive feedback loop there with the developers

Jonny: Look, this is the crux of it. It’s because of the volume of suitability reports and meeting notes being generated. It’s very hard to answer what you’ve just asked accurately. You’d have to analyse every meeting note and report to the nth degree just to get, you know, some data on one single report. So it’s impossible to do that. However, I’ll cover it in two parts, the feedback stage, where we’re in this cycle of developing feedback, you know, sort of iteratively, working with Alan and his team, it was that we were spending a lot of extra time pointing out where it had missed things, where it had hallucinated, and put things in extra so obviously, the the AI guys, and they’ve got an amazing developer, Roshan, who’s the the AI brains behind there. Don’t know how he does what he does, but he keeps doing it. So great. So they take that feedback then and fix the model. So next time it doesn’t do that, or does it a little bit better, so that just gets it better over time. The issue with I see with the false positives, I think you’re right with a high stakes document, like a suitability report, you can’t ever rest on your laurels that you can rely on AI. And this is something I think. Think we’re going to have to probably change how we manage it going forward, because I think there’s a there’s a risk that people get lazy with this. Advisors get lazy with the meeting notes. Oh, the AI is doing a good job. It did a great job on my last one. So I won’t check this one as much. And I think that’s down to you know, me and the management team to just keep instilling that, no, you’re responsible for this. It’s only a tool. You’ve got to check it top and tail it. You’re responsible. That’s ultimately, could be what we need to defend a complaint on, or, you know, fuzzer crawling all over it. And I think, and same goes with the paraplanners, with the suitability report, not get lazy, not just go with what it said, check it, because I think there will always be hallucinations. And that’s actually one of the bugs, if you like, we’re working on with suitability at the moment. Suitability letter at the moment is the paraplanners are giving quite a lot of feedback that the AI is generating more objectives and recommendations than were actually discussed in an ideal world. What we wanted it to do was see if the AI could generate financial planning ideas that hadn’t been picked up by the financial planner. And it can do that because we trialed it with the meeting notes, but then translating that into objectives and recommendations in the suitability report is maybe not what we want. So we might need to sort of take a step back on that particular feature, because it’s just it’s creating more work for the paraplanners than we wanted.

Alasdair: I’m just imagining a an AI is like a little puppy, you know, and you’re trying to teach the puppy to roll over, and the puppy learns, if I roll over, then, then I’m gonna get a get a treat. And the problem is, then you look down at your AI puppy, and it’s rolling all over the floor, constantly waiting for more sweets, more treats. And I feel like that’s that there’s something in the way the AIs are trained. Yeah, you start saying it’s really good when you give us new objectives. Well, then new objectives are gonna come from everywhere and nowhere.

Jonny: Well, yeah, and this is the thing, ultimately, you’ve got to have the audit trail there the objective. There’s no point putting in, well, you can’t put an objective in the suitability report if it wasn’t discussed and agreed with the client in the meeting notes. So it’s all got to tie up. So I think that there is something there that the AI can be like a prompt to remind the financial planner or the paraplanner of things that haven’t been discussed, or maybe shortfalls that they’ve missed, you know, a simple one being, Oh, you didn’t, you know, mention topping up your ISA or, you know, crystallising some capital gains or something. But it’s just finding a balance of that with the output we want.

Heather: Alasdair, I think on the podcast, we’ll have to invite AI in and and we’ll have an AI question on the next on the next podcast. Always invite them to the table, right? I think that’s really fascinating. And that that that concern about how do you keep people on their your toes is such an important topic that we talked about with Pete as well in the last podcast, and I was just thinking about how firms do cyber security checking and fraud detection. And doing, you know, those campaigns to tech. How many people open the phishing email to try and drive training? So it’s, you know, well, we asked the AI to drop easter eggs in there to make sure that those things are caught. It’d be fascinating to see how this develops. I want to, I want to talk about, where next you talked about suitability as sort of something. It sounds like it’s a bit of a work in progress, but I’m really interested in, you know, the examples that you’ve talked about are really making the back end operation much slicker, about driving operational efficiencies, getting humans to focus on what humans do best and where they add value. But do you, you know, do you see that as the main focus for AI development within the firm, or do you also see applications in the front end interface that clients interact with?

Jonny: I think there’s both. I think we’ve got so many ideas and not enough time to do them. There’s, you know, we just keep adding to our sort of wish list of AI. And I think you can break it down by you’ve got potentially client facing uses for this. So that could be, depending how far you want to go down the the AI rabbit hole that’s calling customers, sort of re engagement campaigns. You know, we have a lot of mortgage customers that we don’t always manage to touch base with about protection. Maybe you could use an AI call agent to call them, and just prospect, you’ve got inbound calls coming from clients. Do you use AI to answer that? I phoned up my local pub the other day, and they’ve got an AI agent, like a voice agent, one of the new ones with a decent it’s not like the robot voice you get on call centers. It’s proper AI voice agent was like, this is a little pub in Coddington. Well, they do me that. So I thought that was quite funny. So if they can do it, you know, why aren’t businesses doing it? And you’ve then got the advice process side of it, and that’s what we’re sort of looking at first, because that’s a, you know, probably got the best return on investment, if we can get it right, meeting notes, suitability reports, then you’ve got all the back office functions, provider, data gathering, compliance, you know, I’m sure there’s use cases for the business administration teams. Then there’s things like, what are quite exciting. A bit futuristics are sounding but the the CO pilots the bots that sit along an advisor in the meeting. You know, maybe unbeknownst to the client, and it’s just prompting the advisor listens to the conversation, prompts the advisor, oh, you just mentioned such and such, but have you thought about this? Or this was on the agenda, but you haven’t mentioned it yet? Or, you know, the advisor can ask questions, or they’re going along, or what did the client say earlier? There’s so many really interesting use cases around that, and that tech is already here. The you know, other companies are deploying that in other sectors. And so there’s a real easy read across to financial planners. And another really interesting one is the what Tom was talking about at the conference, about sort of knowledge bases chat bot assistance internally. So instead of an intranet, you know, your staff, if they want to find out what the sickness policy is, they just talk to an AI chat bot, much more efficient and a better use of their time than trolling through our I think we have 41 sort of internal policy documents trying to find the exact bit you need. It could take seconds rather than sort of minutes.

Alasdair: Yeah, this is the bit that’s really that’s really got me thinking, this idea of having some data inside that’s cordoned off, combined with some data outside that gives you that natural language processing at the conference, the event, Tom was talking about this in the context of an assistant for users, but for clients, but the idea that you could effectively just dump all of your existing kind of stories and data and policies and everything, and just Say, right, what is our policy on this? Or, or what’s, what’s going to happen if there’s that is really interesting. And in fact, Heather and I were talking about this yesterday. This is a, this is a today thing. Google have just launched this, this notebook language model where you just give it some stuff, and it just things like creates a podcast of two people speaking about your documents. It is, it’s, there’s that, there’s that quote about, you know, any sufficiently advanced technology looking indistinguishable from magic. This feels like that today, you know. So it’s just really interesting to think how, you know, we’ve got a load of stuff on Confluence. I’m sure Atlassian have got an AI chatbot buried in there somewhere, if I can find it really interesting.

Jonny: Well, it’s like, you know, our compliance policies, they’re all sort of quite disparate. We were actually trying to bring them into, like a master compliance bible, if you like, one place to go and, you know, see all the things you need to be doing on a case. Now, I’m thinking, why bother? Just feed it into a knowledge base, into an AI chatbot system, and let your paraplanners or advisors just use natural language to ask questions. It’s much more efficient, so even little things like that, and this is the thing you can when, and this is what I mentioned at the event. This is why it’s so important to be using AI yourself to understand how it works, to understand what prompts are, how prompts work, to use all these bots and what have you, so you can then think how that might be able to be applied in your business. Because if you don’t have that, that that feel for what AI can do, you won’t see this all the business cases, and there’s just so many, there’s there’s too many, almost to think about 100%.

Heather: 100%, Just for people who were listening, who weren’t at the event. So it’s our AI Lab event. We have room for a few more firms. If you’re interested, you could message Alasdair or me, and it was Tom Whitlock from Clifton who will be a guest on a future episode. We’re still pinning down a date for him, but he’ll share what they’ve been working on at Clifton for their front end or a front end interface for clients that they’re experimenting with. I think we should move to rapid fire questions now, as we as we draw close to wrapping up, we’ve got five questions we ask all of our podcast guests rapid questions. Looking forward to hearing your answers the first one, how can firms get started? You’ve talked a lot about how you got started, but how would you recommend that people get started? But we have a rule no meeting note summaries.

Jonny: I think that there’s one single thing I think people should do, and this, this would apply to anyone, not even in financial services, is pay for a chat GPT license. Ideally, get your boss to pay for a chat GPT license, and start experimenting. Yes, I’ll put my compliance hat here. I. And say, Don’t put client data in or blah, blah, blah. But that’s easy to work around. You don’t need to put client data in to get the benefit of generative AI. So yeah, because once you start using that, understand it, the power of it, you’ll then be in a much better place to move forward about what particular business use cases in your business you want to sort of focus on?

Alasdair: Thank you. Yeah, definitely. You might have answered my final question as well in that, but we’ll get there. So thinking, not necessarily business wise, but something, perhaps a surprising application of AI, or just something where you’ve used AI to make a big difference to your day to day, work or otherwise.

Jonny: I don’t know whether I should say this out loud. These are our favorite kind of examples, because when I said it to friends, they just sort of look at me with that strange look. But is more and more recently, I’ve been using the voice assistant function on chat GPT. So this is where you can chat with generative AI, and I’ve been using it to brainstorm, like projects I’m working on, or ideas or even technical stuff. And you know, I’m talking technical getting quite deep and dirty with REDCap and ifpr and those kind of things, and it understands, yes, it gets some technicality wrong, but it’s enough for me to turn it over in my head when I’m driving to work. So that’s a bit quirky. Talking to AI. I haven’t met anyone else who started doing that yet.

Heather: Well, I do um, not on my way to work, because I take the tube, and that would be, that would be a bit much, I think, on the tube in London, but, um, but no, I 100% agree. And I mean, you mentioned you were a geek. Alasdair certainly is. And, you know, I think this would be a safe space for geeks, right? So maybe, maybe, we’re a minority. But I definitely agree with the chatting with it, because sometimes I’ll type and sometimes I’ll speak, and it depends what I’m working on, but I’ll often be sitting working on my laptop, writing a report and say I’m having trouble wording this. I need to think about this. Find sources that will help me with that and that the it’s collaborating in real time with a colleague that’s very well educated, but without the industry context, and it is incredibly helpful. So I 100% advocate trying that as a different a different way of working with an AI. You’ve mentioned the implement AI podcast, which will definitely put a link to in the show notes. Any other podcast recommendation

Jonny: So if you go back sort of over a year ago, there were hardly any decent AI podcasts. And implement AI was the standout one because it was UK centric. It was business use cases rather than personal use of AI. So it just really hit the nail on the head. I’m surprised it hasn’t got more traction, really, to be honest, I don’t have time to listen to many more. I’ve got other podcasts, and it is something I need to revisit and see if there are any other podcasts out there. But to be honest, the reason I keep recommending that one is because it because it’s a weekly or biweekly podcast, and they really know this space, and it’s business focused, it keeps me on top of things quite easily, without me having to do a lot of reading around the edges, and I just get lots of little prompts and ideas off the back of it to go and investigate. So I’ve had so many good ideas have come out from listening to that does get a bit repetitive, and that, you know, because the end of day, they’re a business, they’re trying to sell their services, but if you read between the lines of what they’re saying and so much to take away from it,

Alasdair: Well, no wonder you’ve got no time to listen to any more podcasts, because you’re too busy chatting to chat GPT, so my this is possibly my favourite question. Heather’s got a great story about a cake, but what is the biggest sort of AI fail you’ve had where you’ve tried to use it something and it just hasn’t worked?

Jonny: Yeah, there’s nothing that there’s nothing that jumps out at me. Everything I’ve tried to do with it, and this isn’t you know me extolling the virtues of AI, it has been able to do it, and if it has come up with a stupid answer, I’ve kind of quickly moved on and just written a prompt in a different way, but I’d love to hear Heather’s at some point,

Heather: I actually wrote a blog about it. I argued with the AI about how you fix ingredients for a cake, and it fought back. It was wrong, but it’s my son’s favorite cake. He still asks for the AI cake. What one thing should people do tomorrow to get started chat?

Jonny: Chat GPT, license and listen to the implement AI podcast.

Heather: Really good suggestions. Thank you, Jonny. Fantastic to hear a little bit more about what you’re working on at LIFT, and it’s great. I mean, you’re an absolute Trailblazer and the thing I love about what you’re doing is you’re not only doing it for the benefit of LIFT, but you’re sharing your experience and what you’ve learned so that other firms can fast track their own path in this area. So thank you so much for that.

Jonny: No problem. Enjoyed it again, enjoyed the event, enjoyed the podcast. Thank you both.

Alasdair: It was great to hear from Jonny just then, we talked quite a bit about our AI Lab event on the 17th of September. Just before Heather and I wrap up with our thoughts on the conversation, I wanted to share some answers to questions that we asked while we were at the AI Lab event. Answers to the first question, what made you sign up for AI Lab in the first place?

Attendee 1: The main reason was with the sheer number of solutions in the noise around AI, I needed to find a way to start to condense that into something that was really useful in the business.

Attendee 2: I think as a firm, we’re quite forward looking. We are very tech led and looking to share best practice, get ideas and talk to other people that think similarly. I think there’s not that many events specifically around tech and AI in our industry. So it felt like a really good opportunity to find like minded firms.

Attendee 3: Well, I mean, AI is such a transformative concept and conversation right now in this space that we just really wanted to have the opportunity to see what some of the advisor firms are doing in this space, what some of the technology houses are doing in this space, really to try and sort of cover more of the waterfront of all the participants, because it’s moving so quickly that you just need to keep listening. You keep need to keep leaning in.

Attendee 4: We’ve been quite aware that AI is on the horizon, and want us to look at ways to implement it into the business. And thought it was a great way to sort of see what other people are doing, sharing best practices, and make sure we went about it in the right way.

Heather: It was great to hear those comments about the event. The feedback has been absolutely fantastic. It’s not just you and I that are buzzing, but lots of people buzzing with ideas, which is fantastic. I thought Jonny was absolutely fantastic. I mean, no surprise, because we’d heard a lot of what he’s doing when he presented at the event this week. I think that his personal curiosity, but his real emotional intelligence about how to position it internally and that journey. I mean, he talked about it like it was an obvious chain of events, but I think that’s actually quite distinct. I mean, he sits on the senior leadership team. He’s a director of the business, but he engaged the executives, engaged the board and and did it in a way that they would respond in the right way, because he didn’t say, you know, we might be able to reduce our risk a bit, interesting, playing right, you know, it was, We can save this much time. I think we can use it and because he has the evidence from time tracking, that’s a fundamental difference right to a lot of firms. So, I think the way that he approached selling that internally is a really good learning and I mean, it’s almost a case study for change management, right? It’s fantastic.

Alasdair: And who’s betting that that was with the help of a few carefully crafted questions to chat GPT as well? You know, how do I persuade our senior leadership team to make those changes? And mysteriously, the chat bot answers. Buy lots more premium licenses for this service.

Heather: Too funny, too funny. You’re a cynic, Alasdair

Alasdair: No, no one’s ever accused of that before. The thing that really gets me about, about Jonny’s approach, and throughout that conversation, you touched on it at the end, actually, there Heather, but, but to expand a little bit as well, this idea of of not just kind of keeping this to himself, because it would be easy to be, you know, first mover advantage. You’re working with probably one of the top three mentioned and kind of hyped financial services specific firms in the UK, specific AI developers in the UK, I should say, and and you’re feeding directly into the into the brain, and you could be forgiven for just not wanting to talk very much about that, you know, to kind of retain that advantage. And he’s so open, and he’s so kind of helpful. I mean, I’ve spoken to him after a couple of things that we’ve been at together, and every time he’s just really willing to sit down and talk about the details and the specifics. And I think that’s why he was so popular at the event, because he really did lay it all out. And as you say, not just in a, not just in a it’s easy to get very deep into the tech and start talking about all the details. But it’s not just that. It’s also a these are the things you need to do with the people as well. So. Yeah, he’s a, he’s a real credit to the development, you know, he is part of the development of AI and financial services in the in the UK, and that’s great.

Heather: Absolutely, something that he sort of alluded to, but it would be difficult for him to talk about was the the impact on the paraplanners. And he talked about people’s worry that they would be replaced by AI. And I was able to of this paraplanner’s assembly event recently. And it was really interesting, because in that room, they said, a year ago at their event, through a show of hands, the majority of people in that room had said, a year ago they thought their job was at risk because of AI and fast forward a year, nobody thought that. And it wasn’t because AI hasn’t delivered what they expected. It’s that now they realised, Oh, now I don’t have to do a suitability report, you know, I have to spend four hours writing a suitability report for a client that wants to top up an ISA. I can spend my time on the really meaty, interesting cases and the places where I can really add value. I’m really applying my brain, you know, it’s, it’s harder work, but it’s much more rewarding work. And so I think that’s, that’s really, you know, that that’s interesting evolution of where we’ve come from to now. And Jonny’s, you know, lived through that, that experience. The other thing that was really interesting that some of the paraplanners mentioned is that every advisor records notes in a different way. And, you know, you run a financial planning practice, and you’ll experience this firsthand, right? But, you know, some people will do a voice recording after a meeting. Some will write up detailed notes. And paraplanners have to produce a consistent output with what they’re given that’s very inconsistent and and so the ability to for the paraplanner to have something consistent to work with that’s complete, and you have that fantastic story about building that query engine into auto which has transformed the way that you you know you can work with your para planners.

Alasdair: Oh, absolutely. I mean, this is, the light bulb moment for me, and we’ve banned people from saying meeting notes, so I feel bad for even mentioning it, but the chat bot built into your meeting notes has been a game changer for us and and we use otter, not affiliated with them, but you know that I’ve used them for years for transcriptions, and they just introduced a chat bot recently. So I thought it was great, because I could structure the notes how I wanted, by saying, Give me a structured summary with these headings that I want for our for our internal notes. But it was a paraplanner that pointed out to me that she could now query the meeting notes, which has got a far better memory about what was said than I have. So if a paraplanner comes to me and says, Oh, you know, did you discuss estate planning, and in what detail I might be like, Hmm, yes, I did. I think maybe once upon a time, you know, and I’ve long forgotten what we’ve talked about, whereas the AI bot says yes, and here are the times during the recording that they were mentioned, and this is what was discussed, and it is just mind blowing the first time you look at it

Heather: Yeah, it still uses ICER instead of Isa, at least it did in the transcription I was reading last week, but, but it’s Getting there right? It’s getting better every day, so I’ve got an action coming out of this. I need to figure out how we can get an AI to come to our next podcast, whether it’s the Zoom AI that can suggest questions so that that’s my action item. Do you have an action coming out of this Alasdair?

Alasdair: I am just absolutely head full of thoughts speaking to Jonny. And I think I’ve said this to you before, but I’m that kind of belligerent person that thinks I need to try and build one of these myself. So my next, my next project, from our discussions about these kind of guardrailed internal help bots, that’s what I want to do. I say, by next week, or probably by next year, we’ll have the H, A, W bot answering all our internal questions.

Heather: Fantastic, coming up next. So we’re releasing these podcasts every other week. Our next guest is going to be Kapil Ram Bhatia from Salesforce, one of our sponsors for the podcast, and that should be really interesting. Salesforce is a, I mean, people think of it as a CRM, but actually that’s a tiny proportion of their business, and they sit behind some of the world’s largest financial advice firms. Have, you know, huge market share within the financial advisor market in the US. And so we’ll have some really, really interesting thoughts for us. Thank you to SS&C, Aviva, Fidelity, Salesforce for their support for AI Lab and for the podcast. Thank you Alistair for CO hosting with me. Thank you to our producer, who pointed out that her initials are AI, which is very appropriate. Artemis Irvin, thank you to you and look forward to not seeing you but sharing thoughts with you in a couple of weeks on our next podcast.

Alasdair: Yeah, catch up with you all soon.