Alicia and Alex outlines practical steps for adopting AI in accounting, noting that 85% of accountants report time savings – especially through workflow automation – with tools like Silverfin. While concerns about data security and job loss persist, AI is largely viewed to enhance, not replace, accountants. The session encourages starting with low-risk tasks to build confidence and aimed to demystify AI, promoting its wider adoption in the industry.
Good afternoon, everyone, and welcome to our webinar today on putting AI into practice. Oh, hello. My name is Alicia, and I head up the customer success team at Silverfin. And I am joined today by the amazing Alex Black from Advalorum. Alex, do you wanna give a bit of an intro to you? Yeah. So, as Alicia said, my name is Alex Black. I’m client services and innovation director at Advalorum. Long title doesn’t really tell you what I do, but I spend about fifty percent of my time working with our clients, strategically from their their side, and I spend about the other fifty percent of my time working in the practice, looking at how we maximize client experience using technology and innovation to do so. Been at Avleron about thirteen years now, so I’ve worked in pretty much every role from trainee up to where I am now and, yeah, seen our journey through that that tech world. Yep. And today, you’re gonna be our AI specialist. You’re gonna carry me through the next hour. Right? I’ll I’ll do my best. Okay. So quick cover off of the agenda and what we’re looking to talk about today. We’re gonna start with looking at the recent Silverfin research project findings, get Alex’s opinion on some of those and share some of those statistics with you. We’re gonna do a quick Silverfin assistant demo. I’m gonna show you how Silverfin are using AI within our own platform. And then Alex is gonna, end the session with some really practical steps on how you can start to implement AI within your firm. And, of course, we’d love a bit of time, for q and a at the end. If any of you have any questions for us, we will find time then as well. So for anyone not familiar with Silverfin, we’re an AI enabled cloud compliant solution that brings together all of your client data. So whether that’s live synced from their bookkeeping systems, pulling that data in, or whether it’s via manually uploaded ledger data, we can handle that as well. We then take all of that data through digital AI enhanced working papers right the way through to accounts production, corporation tax, and reporting. We also have a full communication and review functionality all within our platform. We service over four hundred thousand client files annually within Silverfin, which is great. And our customers typically see around thirty to fifty percent time saving against their old processes. So I’ll touch on the assistant a little bit later. But enough about us for now. Let’s dive straight into it, shall we, Alex, and have a look at the recent all accounted for research project. So this research was undertaken in January twenty twenty five, in partnership with Censuswide. We surveyed around three hundred and fifty mid to senior level accountants on their experiences within our industry. And the main goal really from this project was to understand what matters most to accounting firms when it comes to technology and, of course, when it comes to AI. So I’m really keen to dive into some of these numbers with you and the findings and get your thoughts, Alex. Sounds like a plan. Yeah. I think when I read the report, initially, there’s some bits that quite shocked me, bits I was expecting to see, and there’s some nice surprises in there as well. So, we’re good to go through. Perfect. So the first clear finding, I think, from the report, was that growth is top of the agenda for these accountants. And, actually, this idea that growth really needs talent, but that talent is wanting technology. Right? So forty five percent of those surveyed were said they were severely or significantly affected by the skill shortages. And I think especially when we think about the talent pool that we’re trying to recruit in as accountants, job opportunities in other industries like fintech are actually like, they’re growing at twice the rate of those job roles in accounting, traditional accounting professions. So I don’t think that’s any surprise. But what did really surprise me is that the research also found that on average, firms were investing just nine percent of time into training new staff and actually just eight percent of time into looking at new technology. So I guess, what are your thoughts on that? I think, is AI gonna address those concerns? Is it actually gonna fuel those stats? I think a bit of both depending on on how firms approach this and how partners approach this. I think, you know, we’re talking before it’s it’s that kind of risk versus opportunity position. I think AI is gonna pose a significant risk. You know, it’s the staff base of these firms have got more access to technology in their personal lives than they probably ever had before, at an affordable rate. Therefore, they probably expect the firms to have an even better access to technology because they’ve got more funds and they’ve got bet better economies of scale, sweeping generalization. But, as a, generation, we’re less patient, have less of attention span, probably have greater expectations than we’ve ever had before. And if you take that style of staff member and say, I need you to do a manual day of entry role or I need you to tick this bank over, they’re probably gonna look for another job that’s gonna give them technology and opportunities to to kinda leverage better. Yeah. Gone are the days of the I think when we were talking about this before, the ticky bank rec on a Friday. Spend my day ticking that bank. We don’t don’t get that anymore. And that’s the thing. They’re gonna want no one wants to do that. Right? It’s not that role that that any of us want to do. And if there’s other people out there offering opportunities to not do that, they’re gonna go for that role if I was them. Yeah. And it’s interesting, the graph on the screen at the moment, actually, the percentage of accountants that report still using outdated technology for nonproductive tasks. And I think they’re quite high given where we’re at with technology. And when we say nonproductive tasks, that’s things like copy and pasting, waiting for syncs to load, trying to remote some like, get something off desktop that’s not on cloud. And I think it’s quite interesting that we’re still seeing that. Yeah. I think, again, you’re interested in seeing that higher at that bottom end of the graph, and it’s it’s not a massive distortion. But, I mean, again, demonstrates your eighteen to twenty four year olds are probably reporting it more because they don’t want it. They’re they’re they’re used to not having to do that in their lives than, say, your forty five to fifty four year olds, and they’ve just got higher expectations of what that could look like. Yeah. Definitely. And I think the research did also show, and I think I’ve got it on the next slide, that whilst that, it was just eight percent saying that eight percent of time is looking at new technology, the expectation is that that’s gonna rise. So we can see that actually those survey did expect a seventy nine percent increase in the time they’re spending looking at new technology by twenty twenty eight. So I think that further supports your point, right, that there is appetite for this, and we know we need to be doing it. Yeah. I think, again, opportunity and risk. You know? If that increase in time is potentially a great opportunity for those people to implement in tech, or are they just researching lots and lots of things and then implementing nothing? So is there kind of that that failure to implement, part where they’re almost you got so many options. Like, Netflix, when you scroll for, like, three hours because you can’t find a movie even though you’ve got, like, a thousand options. And I think if we’re not doing this properly, if we’ve not got the right people undertaking this work, then it could become inertia. So I’d probably challenge, it says partners expect. I’d be challenging sort of, is it partners looking at doing this themselves? So are they gonna be delegating this to that talent we were talking about in the last slide and utilizing that talent coming through to undertake this for them? Yeah. No. Definitely. Okay. So for me, a really interesting insight of the survey, and, actually, perhaps one of the highlights, was that contrary to the stereotype, and we could we could dive into those stereotypes a little bit more. But contrary to an accounting stereotype, the data indicated that accountants are actually ahead of the curve of on some of this stuff. So eighty five percent of those surveyed said that AI is already saving them time. And on average, that time saving was reported at about nine percent. And, actually, twenty three percent of those said that that time saving was even higher around eleven to fifteen percent. So I guess it it begs the question, if you’re not using AI yet, why not? Because that’s quite a lot just straight out the box, isn’t it? That’s phenomenal. Right? If if someone said you could have ten percent of your work life back as additional holiday each year, and you bite their hand off. So I think, you know, that’s what we’re looking at here. There’s a part I’m gonna talk about in a minute in the in the survey that talks about, on average, a hundred and seventy six pounds a day of chargeable time could be saved and basically respent. I think it works out about forty eight grand a year, for each staff member in just wasted time. But, also, you know, I did a few numbers, and that could be about forty five days of extra annual leave without any drop in productivity. So if we thought even forty eight thousand pounds forty five days, what would you do with that time? Then it really begs the question of why why are we not using it? How can we use it? What can we use it for? Yeah. And that, leads really nicely onto some of those results, where we saw where we ask those accountants, what are you using AI for? And the top one was workflow automation followed closely by compliance task and communication, which shocked me a little bit. I thought it’d be a little bit different to that. Yeah. Again, I thought communication would probably be at the top, maybe skewed by my own bias as to what I use, AI for. But when I started getting into it, like, this is where I think we’ve got that first what is AI moment or what do people think of AI as? And is this where we’re seeing maybe traditional AI versus gen AI? And I think when we talk about AI these days, my brain instantly goes to Gen AI. I’m thinking of your your copilots, your chat GBT, your Geminis. Whereas maybe, you know, they’re thinking more along the lines of your DEX and and other AI that’s folded into technologies they already use. But it’s phenomenal. I mean, to see that kind of that kind of usage is great. Yeah. Definitely. And I think, I mean, it’s not on the slide, but in the report, it talked, and I know you really liked it, around that invisible vampire of profit. Yeah. And those areas yeah. You loved it. We and those areas where tech where those inefficiencies are being seen, actually, a lot of them, AI will address. So there’s a really worthwhile investment to come from that. Yeah. So, again, for you guys, go read the report. There’s a bit that talks about the invisible vampire profits, which I thought was just a great way to to great word, great way to use to talk about it. But I think there were eight eight tasks, that it talks about kind of being nonproductive inefficient tasks. And when we were looking through them, I went sort of four to five of those. You could solve almost instantaneously with AI, like, basically no effort. And then the other sort of three to four of them, I think, would also do maybe not as easily, but using technology, solve those quite quickly. And so if they are leaching that profit away, if they are taking forty eight thousand pounds a year, per staff member so I think the report talks about, you know, if you have twenty twenty one, twenty staff, that’s a million pounds a year. They’re usually losing in charge of the time. As accountants, we can probably do the ROI on kind of investing in this technology and just kind of give it an hour, give it two hours, kind of play, have a have a go, and see what you can do with it. Yeah. Definitely. And a bit like you, when I was looking at these findings, I expected communication to be the top one, but that’s because that’s the first place that me and my team have started playing around AI. So we introduced, an AI note taker to all of the customer success meetings that we do with customers quite a while ago now, I think, over a year. And it has been a game changer. Just and actually saving fifteen minutes at the end of each call where we’d usually write those tasks to get that time back within the team has been huge for what’s what felt like a really small change. And I bet your notes were better and kind of they didn’t drift off towards the back end and and that that was one of the first things we did as well. We used Copilot because we already were in the Microsoft space. It fit into our environment, and we did our team meeting. So a company wide team meeting, probably a forty, forty five minute meeting. Afterwards, got their AI summary, spent probably five minutes touching it up, and within sort of ten minutes of the meeting ended, we sent out, agenda kind of, summary of what had gone gone on in the meeting, the action points, and kind of what needed to be done before the next. And the response was great because it meant that we had immediately followed up on that meeting, and and people took note and actually did what they need to do off the back of it instead of us getting to the next meeting and going, so where we got them on the action point? They’re like, where are action points? What what were they? Yeah. Yeah. No. Completely agree. So looking across some of the other findings then, we saw that, actually, when it comes to AI, accountants are bucking the trend Across most use cases, the you the research that we completed showed that AI is more deeply embraced by successively older accountants, which, again, I think is contrary to what we’d have expected to see before we did the result. So, for example, just sixteen percent of eighteen to twenty four year olds were using AI for client insights and reporting, and that was compared with fifty percent of those aged fifty five and up. And then for compliance tasks, it was thirty nine percent versus seventy one percent. So that’s a real stark contrast to AI use norms. Like, we we’ve seen it reported by Deloitte, right, which found that sixty two percent of sixteen to thirty four year olds actively use AI compared with just fourteen percent of fifty five to seventy five year olds. So it seems like in our industry, we’re actually going against the grain when it comes to AI. And I I wonder why that is. I’ll be keen to hear your thoughts on on that. Yeah. So I think, again, from our initial meeting, I kind of wracking my brain about it. And I think looking at the Deloitte, I think it was more widespread. Are they using AI or not? And I I would take a stab in the dark that that then their higher number, their sixty two percent is of people that are using it in their personal life or using it in some way outside of profession. And I imagine there’s maybe a a fear, so either a fear of, getting in trouble or maybe there’s policies that tell them they’re not allowed to do their work, or they don’t have the budget. There’s a short the purse strings, budgetary constraints, or they just don’t feel like they have the autonomy to make that decision, and they’re doing what they’re told almost the the kinda old adage of account and what do we do last year? Let’s do the same thing again. Whereas maybe your older staff member is probably more likely to be senior, probably more likely to be in positions of power, and and probably have more autonomy, and more access to funds to use these professionally. So maybe we’re seeing fear almost or or, policy stopping those younger people that have better AI usage. You know, sixty two versus fourteen percent. The argument is they’re probably gonna be better at using it, not using it in the workspace because they don’t feel like they have the autonomy to do so. Yeah. It’s a really interesting point, isn’t it? And, I mean, on the graph there, you can see that that actually it it the trend is across not just that workflow automation compliance task, it’s across all of them, which I find I find really interesting. And I wonder how it will change. Like, as you say, is it a purse strings issue? Is it autonomy issue? As we get more comfortable with AI, I’d be amazed if those numbers don’t change just because the the workforce is predominantly at the younger end. Yeah. And I think, I mean, the the one I wanna pick up on the client insights and reporting. So so a lot of the partners I’ve worked with over the years have got this kind of innate ability to look down a list of quite detailed information and ascertain whether it’s right or it’s wrong, and they’ve got that kind of review ability, which actually is gonna be a superpower in the AI world. So where we’ve got some of those partner more traditionally led people that might not be as tech savvy. They’ve got this superpower of being able to look at information and analyze whether it’s right or wrong, which if we can get AI in the right spaces, it’s gonna be phenomenal. So as much I’d love to see those younger numbers come up, I’d actually love to see the the usage across the board come up because I think there’s something in every generation that’s that’s really gonna add to the benefit. Yeah. Definitely. Definitely. That it leads us really nicely into the next area of the findings, and that was around the concerns in using AI. The biggest of which comes as no surprise. It was data security. I think that is what I expected to see on there because it’s quite new, and and it is a change from what we’re used to. But what do you think about some of the others on there? Yeah. Again, I expected to see data security as the kind of top one. It’s it’s pretty easy one to overcome. So I think, you know, obviously, do your own research, look at your own policies with regards to your your data protection, but a lot of these solutions have enterprise level solutions that don’t retrain the model, that keep your data, ring fenced in your own environment, and and you therefore don’t have a confidentiality issue. So, actually, I think you can probably overcome that big objection straight away. For me, looking across the rest, I actually think we can probably stack fear of job replace, displacement, lack of skills to use, and lack of knowledge and training together. I think they’re ultimately the same thing. And then that becomes the biggest issue, which is, I think, the bigger thing to talk about. If people are scared, they’re gonna be displaced. I think they probably don’t fully understand the technology. And if they don’t think they’ve got the skills or the training to use it, again, I think that comes together and goes, we’ve got a training issue, which harks back to the to the eight percent of time we spoke about at the start of the webinar. Yeah. Well, nine percent for training, eight percent for new tech, but, yeah, it is quite low. And I do, I do agree with you that if you put those three together and they are all really interlinked, it’s a much harder issue to solve than the data security one. Because the data security one will come with confidence and, you know, and processes and our own our own regulation, but that’s a real change of not just a change of approach, but actually underpinning all of your processes, how you train people on what used to be a ticky bank rec, and how you train the the context behind it. Because the I think the slight worry for me is that AI is really great. But when I think back to my time training as an accountant, doing the ticky bank rec, trying to work my way through a massive stage desktop file and understand where all those numbers coming from, the learning came in doing that. And therefore, then as a manager, when I’m reviewing a file, I think you kinda lean on that knowledge of having done it, and you kinda get a we we spoke the other day. We get a bit of an eye for it as to, what you’re looking for and what to expect. But if AI is doing a lot of that for you, how will we how can we still make sure that the new people we’re bringing in get that right knowledge and get the confidence in their knowledge as well outside of what technology is providing them? Yeah. I think that’s probably the the big question. Right? And if we add all those things together, it becomes the biggest issue is how do we train? The fundamental I think one of the comments, in the chat was talking about the the junior should are kinda more focused on improving their practical accounting skills. But the reality is they’re not gonna learn those practical accounting skills the same way you and I did. Right? Because the world’s changed. So we need to find a way to give them that foundation. I mean, I still if if I’m really showing an odd or a t account, kind of map it out in in really manual style. How do we give them that foundation? You know, so they’re not just saying, oh, zero did the double entry. I don’t understand what it is or or whatever that might be. I think that’s where we need to look at our training, look at our education piece, and and find ways to leverage the technology in training rather than trying to almost push it out and go, we don’t need to worry about that till you’re trained. I think you would kind of need to bring it into the conversation and make it all encompassing. Yeah. Definitely. And we, it’s interesting. It comes up again here, actually. We host, every like, twice a year, we host what we call the Silverfin Circle where some of our customers come together, a bit for the customer advisory board and talk about topics. And this was one of the ones that came up in our last meeting a few weeks back around how are we training, how are we recruiting, even down to things like when you’re looking for a new member of staff. When I joined, it was super important that I had a math say level. And, actually, how good I was at maths was probably the bulk of my conversation. But with AI, with technology advancing, I’d argue that, actually, you don’t need to be good at maths to be a really great accountant anymore. There are other skills that are far more important. Those that relationship piece, that client facing element, where technology is taking that legwork where previously it was really important for me to be able to do that myself. Yeah. I I think that you’re gonna see and you you see a lot of that kind of box grid of of what humans are really great at and what we’re not so great at. And then in contrast, what technology or AI, if we wanna replace it, is great at and not great at, and they tend to be kind of equal opposite. So, yes, we’re probably gonna see us leveraging technology far more for calculations and and, you know, deterministic. And we’ve we’ve seen that probably over the last ten years anyway. What we’re now seeing is AI starting to challenge some of those bits or or empower some of those bits that humans traditionally have been way better at. And so it’s going more it’s not a threat necessarily to our position. It’s how do we leverage that to superpower whatever your ability was before and really go that next step. And, you know, when we’re hiring, we definitely look at personality, ability to build relationships before we’re kind of looking at mass stability. You know, you still gotta be there, but it’s not as important as it once was. Yeah. And those skills will transcend through to the next wave of, like, what accounting looks like. Right? Because technology is gonna take some of what used to be like, it already has taken some of that away, but I think the way AI is developing now, it’s only gonna continue to grow there. So this is gonna become a bigger and bigger piece. And, actually, these fears, like, these concerns around the job displacement, that comes down to it’s only gonna be those that aren’t willing to kind of pivot from those traditional skills, I think. There’s also a huge opportunity that comes with, AI and the advances in technology that allow us to focus on the other areas that maybe we haven’t had the time to before. Yeah. Everyone everyone I’ve seen speak speak, especially at the moment. Right? Make ten years down the line, who knows what it’s gonna look like? But at the moment, a human still very much needs to be involved in the process even if it’s a ticky bank crack. So it might just mean if your your life goal was to tick your bank recs to the end, you might just do a hundred times as many now as you did before because you’ve got AI doing a lot of heavy lifting and you’re doing spot checking. Yeah. So I don’t necessarily believe in this massive job displacement bit. Yeah. My bigger fear is how do we get people into account if maths is not the entry? Because that’s definitely how I got it as well. I’m on my you’re good at maths, why don’t you go do accounting? So if that’s verified, how do we get them in the door? And maybe a question for another webinar. Right? Well, look. Of course, we’re seeing AI pop up across all of the platforms. Silverfin’s no different to that. We’re really lucky in that we’ve got an internal AI team. So we bought an AI company a number of years back, and they’ve been integrated into Silverfin as a core part of our team and our offering. And one of the ways that we use AI within our platform is via our assistant tool. So I’m gonna jump in. I’m just gonna show you it for a couple of minutes, and we can chat and see what you think, Alex. So when we come into Silverfin, there’s a number of different places and points at your client journey that, our AI will assist you. The first one is if we add a new period, if that client isn’t live synced with data and we need to start press like, pushing manual data in Silverfin, we absolutely can do it. But the, AI assistant will be there to help prompt you and align those your the columns in your data to what Silverfin’s expecting to see, and it will do some of that mapping for you, which I always find, a bit of a relief because I always get them wrong even as working for Silverfin. And then once we’re in Silverfin, we then have our AI mapping. And this is where it’s Silverfin uses AI to look at your source data. So you can see here my nominal codes. I think these are zero nominal codes. And we need to map it to the Silverfin standard chart so that all that data in Silverfin is the same. We use AI to do that. It will go through a map where it thinks it needs to go in that Silverfin chart of accounts, but it will also flag up to us wherever it’s not really sure. So, actually, AI has done this, but we’re a little bit uncertain. I think it’s other revenue, but it could be sales type one, sales type two. And I think the key thing for me is that, like you’ve said earlier, AI isn’t replacing the need for your expertise and your eyes. It’s helping. So what used to be a really long manual mapping job, actually, we can now really quickly go through, and get that done. We can change that if we want to. We can accept the, report and the decision that the AI has made. The more interesting bit, if we come why is my screen freezing? Always does it to me. The more interesting bit that we see, AI come within the Silverfin platform is with the Silverfin assistant, and you’d have seen us talk about it on LinkedIn. If you’re at the events last year, we had someone running around in a blue morph suit to talk about the assistant. Like, super exciting. But where that fits really nicely is that when we’re working on our client, we have this assistant tab. Now the assistant’s running loads of checks across that data, looking for outliers, looking for anomalies, or wanting to flag points of attention to for your user. We can see that that’s broken down, by working paper. We can see there’s a long list of assistant points that it wants me to take a look at here. First one as an example being, actually, usually where you see computer equipment as a fixed asset, you would also see a code for accumulated depreciation, and you don’t have that. So that we might need to have a little look there. And then we’ve got, for example, dividends. The dividend that’s been distributed is larger than larger than the profits. Other examples are where we’ve got overdrawn loan accounts and things like that. So it’s just being that extra eye to kind of point in the right direction to say, hey. Here’s something we we think you should look at while you’re preparing onto, while you’re preparing your client files. I think these are probably quite reminiscent of some of the bar notes we might have seen based on five in that first review when you were very junior and you get those notes back from your your manager to go, have you looked at your distributor? Or is that Yeah. Yeah. Exactly that. And, like, I always say to customers, like, when what we’re not doing is eliminating the risk of human error. What we are doing is reducing it. Right? Like, those ones where you see the same review point over and over again, those are the checks that now Silfen Assistant is doing and hopefully teaching those users to be looking for it proactively as well rather than waiting for the review point. Are they built in to the system as well? So they’re there, right, for people to use straight away? Yeah. The checks are already in there. Yeah. We don’t set them up. So on this one, for example, we can see in our prepayments tab, the first thing that assistant’s asking me to do is say, actually, you’ve got workings on prepayments in the last in last year’s working paper. Do you wanna roll those forward? And that’s really helpful, particularly with something like HP or fixed asset register. We don’t want someone going into their current year fixed asset register spending ages redoing it when actually they did it last year, and we can just click roll forward and it’s updated for us. Right? So, again, just some of those little sense checks. And then the next one you can see on the screen is saying, actually, I’ve had a look through that ledger. And under advertising and marketing, I can see a transaction that relates to next year, so I think you wanna include that to prepayments. So we can append that, onto our reconciliation. The assistant will then drop that prepayment in with the breakdown. We can check that we’re happy with it and accept those changes. So some of these checks are really gonna save quite a lot of time, give us a really good start point. Another area, which is one that I really enjoy, is starting to look at where we need to have attachments with our templates. So for your VAT reconciliation, for example, we know we’re gonna have to attach those VAT returns. What assistant is gonna do is it’s gonna ask us to upload our VAT returns, which I’m gonna do here. It will then look at those VAT returns, analyze those, and populate our VAT reconciliation with the data as well as attaching those VAT, those VAT returns to the relevant lines. So I click analyze down the bottom. Hopefully, that won’t take too long, and we’ll see that pop in. And this functionality is really great. It’s quite new for us in our AI tool. We also have it across, like, bank loans and HP. So where you’ve got a higher purchase agreement, you can now upload the PDF, and it will extract all of that information and populate your higher purchase template for you. The next one that they’re working on is payroll reconciliation, which I think will go down really well with customers because it’s always the right pain, isn’t it, having to try and type those and tie them up? So you can see here, we’ve got the PDF attachments of the VAT returns, and it’s populated the relevant, the relevant line in our reconciliation with that data for me. So, again, it’s just saving it saves a few minutes, but it’s doing that across every template. It’s a real game changer for where where we’re working to, a chargeable time as well. I think that’s just the time saving there. Right? It’s the accuracy. You know, we’re using Yeah. We’re happy with OCR technology and most other spaces of accounting. We’re using probably similar stuff to to get that accuracy, remove human error. Yeah. Exactly that. And then as we work our way through the client file, like, you can see there’s loads of different checks it does in terms of context, missing data, etcetera. That will carry on working and updating as we continue to work. So it’s a really useful tool and where we’re seeing quite a lot of traction, with Silverfin. We’ve got loads of checks within the platform that we’re running, and we’re also really lucky that we’ve got some really great customers that are using it. They’re using it really well, and giving us feedback on that and helping us to develop the next round of assistant checks. Because that’s the other thing with AI. Like, we don’t wanna it’s not gonna get to a point and stop. It’s always gonna be changing. It’s always gonna be evolving, and we really wanna be at the forefront of that. So really exciting times, Silverfin HQ and some of the customers. We’ve got Russell on the screen here. Some of you will know him. Really enjoying and found within the first few days of using assistance. Started to find some really great, great items that it picked up that maybe that user will have otherwise missed. So, Alex, we’ve taken a look at our research findings. I’ve shown you a bit around how Silverfin are looking at AI and what we’re and what we’re hoping to do and where we’re we’re gonna carry on building out our AI offering. But what I think everyone wants to know is how can we get started with AI if we’re not doing so? Yeah. So Hannah was speaking earlier on. I’m focusing on GenAI here, and I’m gonna lean so actually, I saw a great talk at, Daz this year, and it was, one of the guy the founder of Translucent was on stage, and he was talking about AI. And he kind of had quite a a sharp take to some of the tools you see out there, and he said, you know, a lot of the power buy, and that’s not necessarily concern ourselves for that right now. The ones that say they’re gonna completely replace a human, I don’t believe in that, so let’s not concern ourselves with that. AI assistants, was the the big takeaway. If you’re not already using an AI assistant and we’re talking, Copilot, ChatTBT, Gemini, Brock, and Claude. There’s a bunch out there. They all do this kind of yo yoing thing where they overtake each other, and then the next one releases something new next week and it goes ahead. So pick one. Make sure you’re comfortable with, you know, the user interface. Make sure you’re comfortable with the data protection and stuff, but just pick one. And the biggest take over, just start using it. So, the AI assistant habit. Right? You should be I talk I use mine probably once an hour, if not more. I use mine every single day for tasks throughout the day for various things. And once you get in that habit of using it, challenging it, and building those prompts out, it’s amazing, and it’s so powerful. Yeah. And you know what? When we look at these three practical steps for implementation and you and I were talking about what those steps should be. That first one around building confidence, it is the biggest one. Right? Because I, I like to think I’m quite open to change and open to new technology. In the role that I’m in, I think I’ll probably get sacked if I said otherwise. However, it does still make me nervous, and playing with ChatGPT for things that actually don’t really matter has really given me that, that initial understanding to feel more confident to do it in other areas. So we talked about the meal planning and give me a fitness regime and your meal planner that you’ve that you’ve built within the platform. And I think once you start doing that and you build your confidence in those settings, it’s so much easier to transcend it back into the workplace. Yeah. I think I started trying to get my wife, interested in AI with what I do every single day, and I talked to her about work based ones, and she’s like, doesn’t really understand what I’m talking about. And we were trying to put a meal plan together for the week. I went, I reckon I can do this in Chativity pretty easily. So built a custom GBT, explained what equipment we had in our kitchen, explained what, like, if there are any dietary requirements in the family, I told it. I’ve got three children that are relatively fussy, but I’m not gonna cook them a separate meal. So let’s try and make this, you know, inclusive and build out this this GBT pretty quickly, probably fifteen minutes if that of just typing instructions in and giving it way more detail than I probably needed to. You can tell it what cookbooks you have or what, websites you like to get your recipes from and where you shop so it can create a shopping list afterwards in a format that works for the shop you’re going to. There’s some really cool stuff you can do. But what it did is it got my wife to engage in in AI and and playing with it because before the meal plan for the week, I don’t know if you’re the same as kind of task. We’re like, should we just have the same, like, favorites list that we normally do and not get a lot of variety into that? And instead, you go, actually, give me a five meals. Right? I don’t like that one. Replace it with something that’s veggie, or or, you know, I want something a bit different, and it just makes it so dynamic, and you can start to build it out. And then it’s actually just relatable. You go, actually, I see how that’s practical. Whereas when you go, how are you gonna use it in in business, a lot of people go, not sure what I can do or how to do it. And if you can take a Yeah. Yeah. Definitely. And then that that’s that next point. Right? Identifying those low risk but big reward areas. Once we’ve got that familiarity, we’re feeling a bit more comfortable with it. How can we get a quick win with AI? And I’m we spoke earlier about the communication, the note taker, being an absolute game changer for us. But, actually, when you’re still building confidence, if it if it didn’t work or I wasn’t happy with it, it’s not the end of the world because I’ve just been on the call. I’m not trusting it implicitly. And it and it, again, it just it’s like a a flywheel effect when we’re starting to use AI. And you find that the more you use it, the more confident you get, the more areas you find to use it, you get more confident. It then widens up again. Right? Yeah. And it leads you into step three, which we’ll come on to in a minute, but if you can then if something you could tell colleagues about. So the first one for me, outside the workplace, is something you can tell everyone about, and you’re building your own confidence, and you’re starting to get the ability. The second one, you’re building on that, but it’s how do I start to bring it into professional world? Note taking is an amazing one. We’ve got lots of people asking for, note taking advice. Again, there’s loads of good ones out there. Find what works for your environment. You know, it will depend on whether your Zoom, Teams, Google Meet based. There’s gonna be different ones, but just start using them. And then when you send those meeting notes to all the people involved, they go, this is great. And then they start to get involved. The research assistant for me, another one built a custom research assistant that, can only give me answers from verified sites, so HMRC, International Financial Reporting Standards, CCCA, those kind of sites. Told it where it could go, and then it’ll ask you questions like, you know, what’s FRS one a two starts on this? And it will go and it will it will explain it to me, but then it will also give me a, a link to where I it got that information so I can instantly validate it. So as you’re saying earlier on in terms of its low risk, I’m not gonna trust it implicitly. You can double check its work. But the superpower of the back of that is then you’ve you can then turn it in so many mediums. Right? So you’ve got that information. You’ve got it in the model, and then you go, right, can you draft that up to a client for me using professional language to explain that point? And then, actually, maybe more clients would be interested in this topic than just that one. Can you turn that into a blog for me that I can post on my website so that I can explain this point wider? And then can you write a social media post for me to promote that blog? And you’re gonna have to patch these things up. They’re not gonna be perfect, but they’re gonna be great stars. But even those iterations. Right? Like, can you put that into everyday language? Like, if it’s come straight from the regulation, it is often, like, it is often jargon. Like, can you put it into an easy to understand format that I can send to clients? And then when it spits that back out, it’s like, actually, I don’t like that line. Could you change that line? Actually, I don’t like this bit. Could you change that? Could you like, it’s that ability that takes it beyond just googling. I need to understand this area of their framework. It’s adding that con context around it, which would take you so much longer doing it yourself. Yeah. And I’d and I’d do that. As you just said, I’d do it in the system. Right? So tell it the changes you wanna make. Don’t just copy and paste it to Outlook and do it yourself because you’re training that model, and especially if you build a custom AI, custom GBT, custom Gemini, you’re training it on what you wanted to do next time. You know, you’re giving it the thumbs up, the thumbs down. So give it feedback. Help it make your next answer even better, and you’re gonna get better and better results. And you said you’re gonna save so much time. And I used it the other day to try and explain deferred tax. I’ve done my accountant deferred tax and tried to explain what it was to the client, and they still just weren’t getting it. And as I said, put it into nonjargony language and relate it to a business x. I could have written it. It’d probably take me half an hour to kind of really break it down and get into it. And instead, I spent five minutes touching it up, and it was just a really great use case to try and explain something that’s quite jargony to someone that doesn’t care about that. Yeah. Nice. And then that takes you nicely onto that third point. Right? Like, you’re using it. It’s you’re getting the value. Talk about it. Tell your colleagues. Like, make that firm wide use. Like, we don’t want just one person using this really great tool. We want that to become ingrained across our practice so that we’re all getting the benefit of it. Yeah. And I’ve had so many stories where firms like, there was this one guy who’s just, like, outperforming everyone. When we got into it, he was using, say, a a research assistant on Chativity or something. And you go, instead of just letting him, like, silo that excellence, share that, share those success stories because there’ll be someone else out there that’s got a different strength. They’ve got a different, superpower, and they’ll use it in a different way that he would never have thought of. So for example, I’m terrible at image creation on it. I’m not creative in any way, shape, or form. I’m not given a a professional use case yet, but there are people out there that are far better at that part than I am. So I take information from them where I can. I think we don’t want three, four front runners in your firm that are just sprinting off into the distance. When every anyone’s got a question, they go to them. We want them to bring everyone up to that level so you’re getting that kind of efficiency gain across the whole organization. Yeah. And even things like when you and I were speaking yesterday, you’re telling me about your meal planner. And then, like, oh, I’ve I’ve also looked at it, like, for gardening tips. And I think, oh, I love gardening, and I’m useless at gardening, and I’ve never thought to use it. And it’s things like that. And if you can apply that in a work context, that’s gonna be invaluable for your team. Yeah. And it’s it’s it’s that right? So pick something you’ve got an interest in and test it and see what it can do. Like, I’m terrible at gardening as well. I’d love to have a nice garden. I’ve no idea when I’m supposed to do something. Like, should I cut at the end of the seed? Did did they come out? Are they gonna regrow next year? And it’s just a really quick chain check to tell me this is what you do or this is what is what happens, and then start to apply that to business. And I think, you know, we’re talking now everyone about training. I’ve built some training models. So I built one that was just a really difficult client. So you know those client calls that we all have to have and you kind of your stomach hits the floor when the client every time you say something, they’re like, it’s outrageous, and they’re just objecting to every point built by an obstinate chatbot to try and give people that experience without it being sort of real life stakes. And so they’ve got this back and forth iron messenger with a quite an irate customer that’s upset about something, and it’s a really safe environment for that client to for the staff member to train in without having to wait for an irate client and kinda chuck them in a deep end and go, good luck. That’s a really yeah. It’s and that’s a really good shout. Right? Because, like, you that’s that’s not training you can get. Even if you role play in person, like, it’s quite intense. So to be able to use it for stuff like that is gonna be so helpful. Well, let we’ve we’re reaching the end of time. I did wanna just check for q and a. There are a couple, that I can see up. So the first one is, and you’ve kinda mentioned it already already, Alex, is note taking tools. Like, are there any that you’d recommend? We’ve obviously mentioned it quite a few, and there’s loads of them out there. Right? There are loads of them. I don’t think GBT had one built into it. I think you can give it transcriptions, and it would do it. We’ve used Teams internally so that, Copilot in Teams because we use Teams for our calls anyway. So So it’s already built in. It makes it easier to use. I think you guys use Firefly. Is that right? Yeah. So we implemented Fireflies, and love it. Like, it’s really good. However, we are gonna be testing another one because I think the big thing from this is that we need to stay curious as well because the technology is moving. So I was like, Fireflies is great. Highly recommend. But there are new ones popping up all the time, so we’re also trying to encourage our teams to try new AI, like test it against the other one. So I would my biggest advice would be to try a couple because you’ll find one that fits better for you. Some of the other teams in Silverfin use a different one, for example. So it will come down to your use case as well. Yeah. I saw, I think at at that, there’s a new one that had launched specific for the accountancy world. Yeah. I think it wasn’t necessarily what they’re doing now, but what their plans are in terms of I think that’s the big thing. You said stay curious. These teams are changing so quickly, and it’s what are they gonna do. And if they can fold that into, say, a CRM, which I think these guys were saying they were gonna do, not only have you then got agendas, but you’ve got agendas saved against that specific client in a CRM. How powerful is that? Yeah. Definitely. There’s only one more question. If anyone has got any more questions, I’d encourage you to pop them in. But, Alex, what impact do you think Gen AI will have on the profession, specifically on professional judgment? I think it’s gonna have a massive impact. So I suppose it turns on how you’re defining down that professional judgment part, but, you know, it can look at far greater, parts, the details, information. So one of the great things or one of the tests in early AI was what they called needle in the haystack testing. So they take, say, an hour, two hours of video, and it somewhere in there, there’d be a really obscure comment. Right? It might be, you know, say we were talking, it might be my chef and I said, well, I’ll talk at Tesco’s. And then you go back and you after the hour and a half meeting, someone goes, what shop does Alex shop at? And it searches for all that information and it says, it shops at Tesco’s. And it’s that kind of judgment that it can look at all of the information in the context of a question and deliver that information back, whereas a human has to try and compartmentalize and remember different details. And I think if we if we say take audit, for example, you know, you could be looking at far greater, levels of information to answer one question than a human ever could. So I think you’re still gonna need a human to interpret and definitely to deliver that. But I think if you can leverage AI, you’re gonna be able to have more meaningful conversations with more backup to it as well. It’s not necessarily gonna be, I think you should do this because that’s how I feel. It’s gonna be, I’ve looked at these things, this information. Therefore, I think the best course of action is, and we’re gonna be able to get better advice with more foundation behind it quicker. Yeah. Definitely. And consistency as well. Like, that will be a lot more reliable than professional judgment, which, like, of humans alone, which can vary on any number of circumstances. Yeah. We found it, when was it Xavier, which is now Dex Precision? Yeah. When we first started bringing that in, we did it because clients have terrible bookkeeping records. And when we went back to them and said, your bookkeeping records aren’t good enough. We’re gonna need to charge you an extra fee to fix it. They almost kind of looked at a sketching and going, well, you’re gonna say that because you’re gonna charge me to fix it. When we gave them a Xavier report and it said, here’s your health score. Here are all the issues. If you wanna fix it yourself, there you go. But if you don’t wanna go through all of that, and it was just objective third party thing that said, these are the issues, and then the professional professional comes in to fix it. And I think you’re gonna see more of that across, the profession. Yeah. Definitely. We’ve got one more question that’s popped up, and that is there any tools for in person meetings note takings, which I don’t know. We often have it dialed in to the screen if we’re in a meeting room anyway so we can still use it. Yeah. I’ve done the same. I’ve dialed a Teams meeting in. It does have the unfortunate thing of saying everything’s set by the same person because it’s whose mic it comes through. I’ve seen some tools, so in a past life, I worked in, r and d tax credit space and used to record a lot of meetings. So you can use tools to record and then put them through transcription tools, and that’s pretty good at picking out different voices and giving it, like, different speaker one, speaker two, speaker three, and then you can name them. So I probably look maybe a little bit more old school and go transcription tools and then take that transcription and put it through an AI. I’ve not seen a really good out of the box in person. But if you know of any, please shout because I’m not all over all of them, and it’d be great to see it. Yeah. Definitely. I’ve popped up a quick poll. If any of you would like to know more about Sellofin, please do pop a little link on there, and we’ll be we’ll be more than happy to reach out to you. Thank you for joining. We are gonna be at a number of the events over the summer, so you can see those on screen. And the QR code will take you to a list, not only of the in person events we’re gonna be at, but some of the online bits we’re hosting as well. From me, Alex, it’s just a really big thank you for joining me and for sharing all of your wisdom. It’s been super valuable for me, and I’m sure everyone on the call as well. Thank Thank you for having me. It’s been a pleasure. Lovely. Thank you all for joining. Enjoy the rest of your day. That’s great, guys.